├── .editorconfig ├── .flake8 ├── .github └── workflows │ ├── python.yml │ └── release.yml ├── .gitignore ├── Dockerfile ├── LICENSE ├── MANIFEST.in ├── README.md ├── kafka_connect_healthcheck ├── __init__.py ├── handler.py ├── health.py ├── helpers.py ├── main.py ├── parser.py └── version.py ├── requirements-dev.txt ├── scripts ├── build-docker.sh ├── lint.sh ├── test-coverage.sh └── test.sh ├── setup.cfg ├── setup.py ├── tests ├── __init__.py ├── conftest.py ├── context.py ├── data │ ├── expected │ │ ├── 0-healthcheck-server-healthy.json │ │ ├── 1-healthy.json │ │ ├── 10-unhealthy-multiple-connectors.json │ │ ├── 11-healthy-no-connectors.json │ │ ├── 12-unhealthy-task-with-trace.json │ │ ├── 13-unhealthy-broker-connection.json │ │ ├── 14-basic-auth.json │ │ ├── 15-unhealthy-threshold.json │ │ ├── 16-healthy-threshold.json │ │ ├── 17-healthy-container-connector.json │ │ ├── 18-healthy-container-task.json │ │ ├── 19-unhealthy-container-task.json │ │ ├── 2-unhealthy.json │ │ ├── 3-unhealthy-error.json │ │ ├── 4-healthy-worker-id-correct.json │ │ ├── 5-healthy-worker-id-unused.json │ │ ├── 6-healthy-worker-id-with-other-workers-failing.json │ │ ├── 7-unhealthy-worker-id-with-other-workers-healthy.json │ │ ├── 8-healthy-multiple-tasks.json │ │ └── 9-healthy-multiple-connectors.json │ └── mocks │ │ ├── 1-healthy-connector-local-file-source.json │ │ ├── 1-healthy-connectors.json │ │ ├── 10-unhealthy-multiple-connectors-connector-s3-sink.json │ │ ├── 10-unhealthy-multiple-connectors-connector-s3-source.json │ │ ├── 10-unhealthy-multiple-connectors-connectors.json │ │ ├── 11-healthy-no-connectors-connectors.json │ │ ├── 12-unhealthy-task-with-trace-connector-s3-sink.json │ │ ├── 12-unhealthy-task-with-trace-connectors.json │ │ ├── 13-unhealthy-broker-connection-connector-s3-sink.json │ │ ├── 13-unhealthy-broker-connection-connectors.json │ │ ├── 14-basic-auth-connector-local-file-source.json │ │ ├── 14-basic-auth-connectors.json │ │ ├── 15-unhealthy-threshold-connector-s3-sink.json │ │ ├── 15-unhealthy-threshold-connector-s3-source.json │ │ ├── 15-unhealthy-threshold-connectors.json │ │ ├── 16-healthy-threshold-connector-s3-sink.json │ │ ├── 16-healthy-threshold-connector-s3-source.json │ │ ├── 16-healthy-threshold-connectors.json │ │ ├── 17-healthy-container-connector-connector-local-file-source.json │ │ ├── 17-healthy-container-connector-connectors.json │ │ ├── 18-healthy-container-task-connector-local-file-source.json │ │ ├── 18-healthy-container-task-connectors.json │ │ ├── 19-unhealthy-container-task-connector-local-file-source.json │ │ ├── 19-unhealthy-container-task-connectors.json │ │ ├── 2-unhealthy-connector-local-file-source.json │ │ ├── 2-unhealthy-connectors.json │ │ ├── 4-healthy-worker-id-correct-connector-jdbc-sink.json │ │ ├── 4-healthy-worker-id-correct-connectors.json │ │ ├── 5-healthy-worker-id-unused-connector-jdbc-sink.json │ │ ├── 5-healthy-worker-id-unused-connectors.json │ │ ├── 6-healthy-worker-id-with-other-workers-failing-connector-jdbc-sink.json │ │ ├── 6-healthy-worker-id-with-other-workers-failing-connectors.json │ │ ├── 7-unhealthy-worker-id-with-other-workers-healthy-connector-jdbc-sink.json │ │ ├── 7-unhealthy-worker-id-with-other-workers-healthy-connectors.json │ │ ├── 8-healthy-multiple-tasks-connector-jdbc-source.json │ │ ├── 8-healthy-multiple-tasks-connectors.json │ │ ├── 9-healthy-multiple-connectors-connector-s3-sink.json │ │ ├── 9-healthy-multiple-connectors-connector-s3-source.json │ │ ├── 9-healthy-multiple-connectors-connectors.json │ │ └── healthy-connector-details.json ├── mocks.py ├── test_server.py └── test_version.py └── tox.ini /.editorconfig: -------------------------------------------------------------------------------- 1 | root = true 2 | 3 | [*] 4 | charset = utf-8 5 | end_of_line = lf 6 | insert_final_newline = true 7 | trim_trailing_whitespace = true 8 | 9 | [{*.py, *.json}] 10 | indent_style = space 11 | indent_size = 4 12 | -------------------------------------------------------------------------------- /.flake8: -------------------------------------------------------------------------------- 1 | select = Q0 2 | -------------------------------------------------------------------------------- /.github/workflows/python.yml: -------------------------------------------------------------------------------- 1 | name: Python CI 2 | 3 | on: 4 | push: 5 | branches: 6 | - '**' 7 | tags-ignore: 8 | - '**' 9 | 10 | jobs: 11 | build: 12 | runs-on: ubuntu-latest 13 | strategy: 14 | max-parallel: 4 15 | matrix: 16 | python-version: [3.5, 3.6, 3.7, 3.8] 17 | 18 | steps: 19 | - uses: actions/checkout@v1 20 | - name: Set up Python ${{ matrix.python-version }} 21 | uses: actions/setup-python@v2 22 | with: 23 | python-version: ${{ matrix.python-version }} 24 | - name: Install dependencies 25 | run: | 26 | python -m pip install --upgrade pip 27 | python -m pip install -r requirements-dev.txt 28 | python -m pip install -e . 29 | pip install tox tox-gh-actions 30 | - name: Test with tox 31 | run: tox 32 | -------------------------------------------------------------------------------- /.github/workflows/release.yml: -------------------------------------------------------------------------------- 1 | on: 2 | push: 3 | tags: 4 | - '*' 5 | 6 | name: Release 7 | 8 | jobs: 9 | build: 10 | name: Build Release 11 | runs-on: ubuntu-latest 12 | steps: 13 | - uses: actions/checkout@v1 14 | 15 | - name: Create Release 16 | id: create_release 17 | uses: actions/create-release@v1.0.0 18 | env: 19 | GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }} 20 | with: 21 | tag_name: ${{ github.ref }} 22 | release_name: Release ${{ github.ref }} 23 | draft: true 24 | prerelease: false 25 | -------------------------------------------------------------------------------- /.gitignore: -------------------------------------------------------------------------------- 1 | .idea/ 2 | venv/ 3 | dist/ 4 | build/ 5 | htmlcov/ 6 | .pytest_cache/ 7 | /*.egg-info 8 | __pycache__ 9 | *.iml 10 | *.pyc 11 | .coverage* 12 | .tox/ 13 | examples/ 14 | -------------------------------------------------------------------------------- /Dockerfile: -------------------------------------------------------------------------------- 1 | FROM python:3.7-alpine 2 | 3 | COPY setup.py README.md LICENSE /kafka-connect-healthcheck/ 4 | COPY kafka_connect_healthcheck/ /kafka-connect-healthcheck/kafka_connect_healthcheck/ 5 | 6 | RUN cd /kafka-connect-healthcheck && pip3 install -e . 7 | 8 | CMD ["kafka-connect-healthcheck"] 9 | -------------------------------------------------------------------------------- /LICENSE: -------------------------------------------------------------------------------- 1 | Apache License 2 | Version 2.0, January 2004 3 | http://www.apache.org/licenses/ 4 | 5 | TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION 6 | 7 | 1. Definitions. 8 | 9 | "License" shall mean the terms and conditions for use, reproduction, 10 | and distribution as defined by Sections 1 through 9 of this document. 11 | 12 | "Licensor" shall mean the copyright owner or entity authorized by 13 | the copyright owner that is granting the License. 14 | 15 | "Legal Entity" shall mean the union of the acting entity and all 16 | other entities that control, are controlled by, or are under common 17 | control with that entity. For the purposes of this definition, 18 | "control" means (i) the power, direct or indirect, to cause the 19 | direction or management of such entity, whether by contract or 20 | otherwise, or (ii) ownership of fifty percent (50%) or more of the 21 | outstanding shares, or (iii) beneficial ownership of such entity. 22 | 23 | "You" (or "Your") shall mean an individual or Legal Entity 24 | exercising permissions granted by this License. 25 | 26 | "Source" form shall mean the preferred form for making modifications, 27 | including but not limited to software source code, documentation 28 | source, and configuration files. 29 | 30 | "Object" form shall mean any form resulting from mechanical 31 | transformation or translation of a Source form, including but 32 | not limited to compiled object code, generated documentation, 33 | and conversions to other media types. 34 | 35 | "Work" shall mean the work of authorship, whether in Source or 36 | Object form, made available under the License, as indicated by a 37 | copyright notice that is included in or attached to the work 38 | (an example is provided in the Appendix below). 39 | 40 | "Derivative Works" shall mean any work, whether in Source or Object 41 | form, that is based on (or derived from) the Work and for which the 42 | editorial revisions, annotations, elaborations, or other modifications 43 | represent, as a whole, an original work of authorship. For the purposes 44 | of this License, Derivative Works shall not include works that remain 45 | separable from, or merely link (or bind by name) to the interfaces of, 46 | the Work and Derivative Works thereof. 47 | 48 | "Contribution" shall mean any work of authorship, including 49 | the original version of the Work and any modifications or additions 50 | to that Work or Derivative Works thereof, that is intentionally 51 | submitted to Licensor for inclusion in the Work by the copyright owner 52 | or by an individual or Legal Entity authorized to submit on behalf of 53 | the copyright owner. For the purposes of this definition, "submitted" 54 | means any form of electronic, verbal, or written communication sent 55 | to the Licensor or its representatives, including but not limited to 56 | communication on electronic mailing lists, source code control systems, 57 | and issue tracking systems that are managed by, or on behalf of, the 58 | Licensor for the purpose of discussing and improving the Work, but 59 | excluding communication that is conspicuously marked or otherwise 60 | designated in writing by the copyright owner as "Not a Contribution." 61 | 62 | "Contributor" shall mean Licensor and any individual or Legal Entity 63 | on behalf of whom a Contribution has been received by Licensor and 64 | subsequently incorporated within the Work. 65 | 66 | 2. Grant of Copyright License. Subject to the terms and conditions of 67 | this License, each Contributor hereby grants to You a perpetual, 68 | worldwide, non-exclusive, no-charge, royalty-free, irrevocable 69 | copyright license to reproduce, prepare Derivative Works of, 70 | publicly display, publicly perform, sublicense, and distribute the 71 | Work and such Derivative Works in Source or Object form. 72 | 73 | 3. Grant of Patent License. Subject to the terms and conditions of 74 | this License, each Contributor hereby grants to You a perpetual, 75 | worldwide, non-exclusive, no-charge, royalty-free, irrevocable 76 | (except as stated in this section) patent license to make, have made, 77 | use, offer to sell, sell, import, and otherwise transfer the Work, 78 | where such license applies only to those patent claims licensable 79 | by such Contributor that are necessarily infringed by their 80 | Contribution(s) alone or by combination of their Contribution(s) 81 | with the Work to which such Contribution(s) was submitted. If You 82 | institute patent litigation against any entity (including a 83 | cross-claim or counterclaim in a lawsuit) alleging that the Work 84 | or a Contribution incorporated within the Work constitutes direct 85 | or contributory patent infringement, then any patent licenses 86 | granted to You under this License for that Work shall terminate 87 | as of the date such litigation is filed. 88 | 89 | 4. Redistribution. You may reproduce and distribute copies of the 90 | Work or Derivative Works thereof in any medium, with or without 91 | modifications, and in Source or Object form, provided that You 92 | meet the following conditions: 93 | 94 | (a) You must give any other recipients of the Work or 95 | Derivative Works a copy of this License; and 96 | 97 | (b) You must cause any modified files to carry prominent notices 98 | stating that You changed the files; and 99 | 100 | (c) You must retain, in the Source form of any Derivative Works 101 | that You distribute, all copyright, patent, trademark, and 102 | attribution notices from the Source form of the Work, 103 | excluding those notices that do not pertain to any part of 104 | the Derivative Works; and 105 | 106 | (d) If the Work includes a "NOTICE" text file as part of its 107 | distribution, then any Derivative Works that You distribute must 108 | include a readable copy of the attribution notices contained 109 | within such NOTICE file, excluding those notices that do not 110 | pertain to any part of the Derivative Works, in at least one 111 | of the following places: within a NOTICE text file distributed 112 | as part of the Derivative Works; within the Source form or 113 | documentation, if provided along with the Derivative Works; or, 114 | within a display generated by the Derivative Works, if and 115 | wherever such third-party notices normally appear. The contents 116 | of the NOTICE file are for informational purposes only and 117 | do not modify the License. You may add Your own attribution 118 | notices within Derivative Works that You distribute, alongside 119 | or as an addendum to the NOTICE text from the Work, provided 120 | that such additional attribution notices cannot be construed 121 | as modifying the License. 122 | 123 | You may add Your own copyright statement to Your modifications and 124 | may provide additional or different license terms and conditions 125 | for use, reproduction, or distribution of Your modifications, or 126 | for any such Derivative Works as a whole, provided Your use, 127 | reproduction, and distribution of the Work otherwise complies with 128 | the conditions stated in this License. 129 | 130 | 5. Submission of Contributions. Unless You explicitly state otherwise, 131 | any Contribution intentionally submitted for inclusion in the Work 132 | by You to the Licensor shall be under the terms and conditions of 133 | this License, without any additional terms or conditions. 134 | Notwithstanding the above, nothing herein shall supersede or modify 135 | the terms of any separate license agreement you may have executed 136 | with Licensor regarding such Contributions. 137 | 138 | 6. Trademarks. This License does not grant permission to use the trade 139 | names, trademarks, service marks, or product names of the Licensor, 140 | except as required for reasonable and customary use in describing the 141 | origin of the Work and reproducing the content of the NOTICE file. 142 | 143 | 7. Disclaimer of Warranty. Unless required by applicable law or 144 | agreed to in writing, Licensor provides the Work (and each 145 | Contributor provides its Contributions) on an "AS IS" BASIS, 146 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or 147 | implied, including, without limitation, any warranties or conditions 148 | of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A 149 | PARTICULAR PURPOSE. You are solely responsible for determining the 150 | appropriateness of using or redistributing the Work and assume any 151 | risks associated with Your exercise of permissions under this License. 152 | 153 | 8. Limitation of Liability. In no event and under no legal theory, 154 | whether in tort (including negligence), contract, or otherwise, 155 | unless required by applicable law (such as deliberate and grossly 156 | negligent acts) or agreed to in writing, shall any Contributor be 157 | liable to You for damages, including any direct, indirect, special, 158 | incidental, or consequential damages of any character arising as a 159 | result of this License or out of the use or inability to use the 160 | Work (including but not limited to damages for loss of goodwill, 161 | work stoppage, computer failure or malfunction, or any and all 162 | other commercial damages or losses), even if such Contributor 163 | has been advised of the possibility of such damages. 164 | 165 | 9. Accepting Warranty or Additional Liability. While redistributing 166 | the Work or Derivative Works thereof, You may choose to offer, 167 | and charge a fee for, acceptance of support, warranty, indemnity, 168 | or other liability obligations and/or rights consistent with this 169 | License. However, in accepting such obligations, You may act only 170 | on Your own behalf and on Your sole responsibility, not on behalf 171 | of any other Contributor, and only if You agree to indemnify, 172 | defend, and hold each Contributor harmless for any liability 173 | incurred by, or claims asserted against, such Contributor by reason 174 | of your accepting any such warranty or additional liability. 175 | 176 | END OF TERMS AND CONDITIONS 177 | 178 | APPENDIX: How to apply the Apache License to your work. 179 | 180 | To apply the Apache License to your work, attach the following 181 | boilerplate notice, with the fields enclosed by brackets "[]" 182 | replaced with your own identifying information. (Don't include 183 | the brackets!) The text should be enclosed in the appropriate 184 | comment syntax for the file format. We also recommend that a 185 | file or class name and description of purpose be included on the 186 | same "printed page" as the copyright notice for easier 187 | identification within third-party archives. 188 | 189 | Copyright 2019 Shawn Seymour 190 | 191 | Licensed under the Apache License, Version 2.0 (the "License"); 192 | you may not use this file except in compliance with the License. 193 | You may obtain a copy of the License at 194 | 195 | http://www.apache.org/licenses/LICENSE-2.0 196 | 197 | Unless required by applicable law or agreed to in writing, software 198 | distributed under the License is distributed on an "AS IS" BASIS, 199 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 200 | See the License for the specific language governing permissions and 201 | limitations under the License. -------------------------------------------------------------------------------- /MANIFEST.in: -------------------------------------------------------------------------------- 1 | include LICENSE 2 | include README.md 3 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # kafka-connect-healthcheck 2 | 3 | ![Python CI](https://github.com/devshawn/kafka-connect-healthcheck/workflows/Python%20CI/badge.svg) [![Docker Cloud Build Status](https://img.shields.io/docker/cloud/build/devshawn/kafka-connect-healthcheck.svg)](https://hub.docker.com/r/devshawn/kafka-connect-healthcheck) ![PyPI](https://img.shields.io/pypi/v/kafka-connect-healthcheck.svg?color=blue) ![PyPI - Python Version](https://img.shields.io/pypi/pyversions/kafka-connect-healthcheck.svg) [![License](https://img.shields.io/badge/License-Apache%202.0-blue.svg)](LICENSE) 4 | 5 | 6 | A simple healthcheck wrapper to monitor Kafka Connect. 7 | 8 | 9 |

10 | 11 |

12 | 13 | Kafka Connect Healthcheck is a server that wraps the Kafka Connect API and provides a singular API endpoint to determine the health of a Kafka Connect instance. This can be used to alert or take action on unhealthy connectors and tasks. 14 | 15 | This can be used in numerous ways. It can sit as a standalone service for monitoring purposes, it can be used as a sidecar container to mark Kafka Connect workers as unhealthy in Kubernetes, or it can be used to provide logs of when connectors/tasks failed and reasons for their failures. 16 | 17 | By default, the root endpoint `/` will return `200 OK` healthy if all connectors and tasks are in a state other than `FAILED`. It will return `503 Service Unavailable` if any connector or tasks are in a `FAILED` state. 18 | 19 | ## Usage 20 | 21 | Kafka Connect Healthcheck can be installed as a command-line tool through `pip` or it can be used as a standalone Docker container. It could also be installed as a part of a custom Kafka Connect docker image. 22 | 23 | ### Command-Line 24 | To use `kafka-connect-healthcheck` from the command-line, you must have `python` and `pip` installed. Currently, only Python 3 is supported. 25 | 26 | You can install `kafka-connect-healthcheck` via pip: 27 | 28 | ```bash 29 | pip install kafka-connect-healthcheck 30 | ``` 31 | 32 | To start the healthcheck server, run: 33 | 34 | ```bash 35 | kafka-connect-healthcheck 36 | ``` 37 | 38 | The server will now be running on [localhost:18083][localhost]. 39 | 40 | ### Docker 41 | The `kafka-connect-healthcheck` image can be found on Docker Hub. 42 | 43 | You can pull down the latest image by running: 44 | 45 | ```bash 46 | docker pull devshawn/kafka-connect-healthcheck 47 | ``` 48 | 49 | To start the healthcheck server, run: 50 | 51 | ```bash 52 | docker run --rm -it -p 18083:18083 devshawn/kafka-connect-healthcheck 53 | ``` 54 | 55 | The server will now be running on [localhost:18083][localhost]. 56 | 57 | ## Configuration 58 | Kafka Connect Healthcheck can be configured via command-line arguments or by environment variables. 59 | 60 | #### Port 61 | The port for the `kafka-connect-healthcheck` API. 62 | 63 | | Usage | Value | 64 | |-----------------------|--------------------| 65 | | Environment Variable | `HEALTHCHECK_PORT` | 66 | | Command-Line Argument | `--port` | 67 | | Default Value | `18083` | 68 | 69 | #### Connect URL 70 | The full URL of the Kafka Connect REST API. This is used to determine the health of the connect instance. 71 | 72 | | Usage | Value | 73 | |-----------------------|---------------------------| 74 | | Environment Variable | `HEALTHCHECK_CONNECT_URL` | 75 | | Command-Line Argument | `--connect-url` | 76 | | Default Value | `http://localhost:8083` | 77 | 78 | #### Connect Worker ID 79 | The worker ID to monitor (usually the IP address of the connect worker). If none is set, all workers will be monitored and any failure will result in an unhealthy response. 80 | 81 | | Usage | Value | 82 | |-----------------------|---------------------------------| 83 | | Environment Variable | `HEALTHCHECK_CONNECT_WORKER_ID` | 84 | | Command-Line Argument | `--connect-worker-id` | 85 | | Default Value | None (all workers monitored) | 86 | 87 | **Note**: It is highly recommended to run an instance of the healthcheck for each worker if you're planning to restart containers based on the health. 88 | 89 | #### Considered Containers 90 | A comma-separated list of which type of kafka connect container to be considered in the healthcheck calculation. 91 | 92 | | Usage | Value | 93 | |-----------------------|---------------------------------------------| 94 | | Environment Variable | `HEALTHCHECK_CONSIDERED_CONTAINERS` | 95 | | Command-Line Argument | `--considered-containers` | 96 | | Default Value | `CONNECTOR,TASK` | 97 | | Valid Values | `CONNECTOR`, `TASK` | 98 | 99 | #### Unhealthy States 100 | A comma-separated list of connector and tasks states to be marked as unhealthy. 101 | 102 | | Usage | Value | 103 | |-----------------------|---------------------------------------------| 104 | | Environment Variable | `HEALTHCHECK_UNHEALTHY_STATES` | 105 | | Command-Line Argument | `--unhealthy-states` | 106 | | Default Value | `FAILED` | 107 | | Valid Values | `FAILED`, `PAUSED`, `UNASSIGNED`, `RUNNING` | 108 | 109 | **Note**: It's recommended to keep this defaulted to `FAILED`, but paused connectors or tasks can be marked as unhealthy by passing `FAILED,PAUSED`. 110 | 111 | #### Failure Threshold Percentage 112 | A number between 1 and 100. If set, this is the percentage of connectors that must fail for the healthcheck to fail. 113 | 114 | | Usage | Value | 115 | |-----------------------|---------------------------------------------| 116 | | Environment Variable | `HEALTHCHECK_FAILURE_THRESHOLD_PERCENTAGE` | 117 | | Command-Line Argument | `--failure-threshold-percentage` | 118 | | Default Value | `0` | 119 | | Valid Values | 1 to 100 | 120 | 121 | By default, **any** failures will cause the healthcheck to fail. 122 | 123 | #### Log Level 124 | The level of logs to be shown by the application. 125 | 126 | | Usage | Value | 127 | |-----------------------|---------------------------------------------| 128 | | Environment Variable | `HEALTHCHECK_LOG_LEVEL` | 129 | | Command-Line Argument | `--log-level` | 130 | | Default Value | `INFO` | 131 | | Valid Values | `DEBUG`, `INFO`, `WARNING`, `ERROR` | 132 | 133 | All healthy connector and task statuses are logged at `INFO`. Unhealthy ones are logged at `WARNING`. Any communication or HTTP errors are logged at `ERROR`. 134 | 135 | ## API 136 | The server provides a very simple HTTP API which can be used for liveness probes and monitoring alerts. We expose two endpoints: 137 | 138 | #### `GET /` 139 | Get the current health status of the Kafka Connect system. This could be used as a sidecar to determine the health of each Kafka Connect worker and their associated connectors and tasks. 140 | 141 | **Example Request** 142 | ```bash 143 | curl http://localhost:18083 144 | ``` 145 | 146 | **Example Healthy Response** 147 | 148 | 200 OK 149 | ```json 150 | { 151 | "failures": [], 152 | "failure_states": [ 153 | "FAILED" 154 | ], 155 | "healthy": true 156 | } 157 | ``` 158 | 159 | **Example Unhealthy Response** 160 | 161 | 503 Service Unavailable 162 | ```json 163 | { 164 | "failures": [ 165 | { 166 | "type": "connector", 167 | "connector": "jdbc-source", 168 | "state": "FAILED", 169 | "worker_id": "127.0.0.1:8083" 170 | }, 171 | { 172 | "type": "task", 173 | "connector": "jdbc-source", 174 | "id": 0, 175 | "state": "FAILED", 176 | "worker_id": "127.0.0.1:8083", 177 | "trace": "..." 178 | } 179 | ], 180 | "failure_states": [ 181 | "FAILED" 182 | ], 183 | "healthy": false 184 | } 185 | ``` 186 | 187 | #### `GET /ping` 188 | Get the current health status of the healthcheck server. This will always be successful as long as the server is still able to serve requests. This can be used as a ready or liveness probe in Kubernetes. 189 | 190 | **Example Request** 191 | ```bash 192 | curl http://localhost:18083/ping 193 | ``` 194 | 195 | **Example Response** 196 | 197 | 200 OK 198 | ```json 199 | { 200 | "status": "UP" 201 | } 202 | ``` 203 | 204 | ## License 205 | Copyright (c) 2019 Shawn Seymour. 206 | 207 | Licensed under the [Apache 2.0 license][license]. 208 | 209 | [localhost]: http://localhost:18083 210 | [license]: LICENSE 211 | -------------------------------------------------------------------------------- /kafka_connect_healthcheck/__init__.py: -------------------------------------------------------------------------------- 1 | # -*- coding: utf-8 -*- 2 | # 3 | # Copyright 2019 Shawn Seymour. All Rights Reserved. 4 | # 5 | # Licensed under the Apache License, Version 2.0 (the "License"). You 6 | # may not use this file except in compliance with the License. A copy of 7 | # the License is located at 8 | # 9 | # http://www.apache.org/licenses/LICENSE-2.0 10 | # 11 | # or in the "license" file accompanying this file. This file is 12 | # distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF 13 | # ANY KIND, either express or implied. See the License for the specific 14 | # language governing permissions and limitations under the License. 15 | 16 | from kafka_connect_healthcheck import handler 17 | from kafka_connect_healthcheck import health 18 | from kafka_connect_healthcheck import helpers 19 | from kafka_connect_healthcheck import main 20 | from kafka_connect_healthcheck import version 21 | 22 | name = "kafka_connect_healthcheck" 23 | 24 | __all__ = [ 25 | "handler", 26 | "health", 27 | "helpers", 28 | "main", 29 | "version" 30 | ] 31 | -------------------------------------------------------------------------------- /kafka_connect_healthcheck/handler.py: -------------------------------------------------------------------------------- 1 | # -*- coding: utf-8 -*- 2 | # 3 | # Copyright 2019 Shawn Seymour. All Rights Reserved. 4 | # 5 | # Licensed under the Apache License, Version 2.0 (the "License"). You 6 | # may not use this file except in compliance with the License. A copy of 7 | # the License is located at 8 | # 9 | # http://www.apache.org/licenses/LICENSE-2.0 10 | # 11 | # or in the "license" file accompanying this file. This file is 12 | # distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF 13 | # ANY KIND, either express or implied. See the License for the specific 14 | # language governing permissions and limitations under the License. 15 | 16 | import json 17 | 18 | from http.server import BaseHTTPRequestHandler 19 | 20 | 21 | class RequestHandler(BaseHTTPRequestHandler): 22 | 23 | def __init__(self, health, *args, **kwargs): 24 | self.health = health 25 | super().__init__(*args, **kwargs) 26 | 27 | def do_HEAD(self): 28 | self.send_response(200) 29 | self.send_header("Content-type", "text/html") 30 | self.end_headers() 31 | 32 | def do_GET(self): 33 | if self.path in ["/"]: 34 | payload = self.health.get_health_result() 35 | payload_json = json.dumps(payload) 36 | status = 200 if payload["healthy"] else 503 37 | self.respond(status, payload_json) 38 | elif self.path in ["/ping"]: 39 | payload_json = json.dumps({"status": "UP"}) 40 | self.respond(200, payload_json) 41 | else: 42 | self.respond(404, "") 43 | 44 | def log_message(self, format, *args): 45 | return 46 | 47 | def handle_http(self, status_code, payload): 48 | self.send_response(status_code) 49 | self.send_header("Content-type", "application/json") 50 | self.end_headers() 51 | return bytes(payload, "UTF-8") 52 | 53 | def respond(self, status, payload): 54 | response = self.handle_http(status, payload) 55 | self.wfile.write(response) 56 | -------------------------------------------------------------------------------- /kafka_connect_healthcheck/health.py: -------------------------------------------------------------------------------- 1 | # -*- coding: utf-8 -*- 2 | # 3 | # Copyright 2019 Shawn Seymour. All Rights Reserved. 4 | # 5 | # Licensed under the Apache License, Version 2.0 (the "License"). You 6 | # may not use this file except in compliance with the License. A copy of 7 | # the License is located at 8 | # 9 | # http://www.apache.org/licenses/LICENSE-2.0 10 | # 11 | # or in the "license" file accompanying this file. This file is 12 | # distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF 13 | # ANY KIND, either express or implied. See the License for the specific 14 | # language governing permissions and limitations under the License. 15 | 16 | import logging 17 | 18 | import requests 19 | 20 | from kafka_connect_healthcheck import helpers 21 | 22 | 23 | class Health: 24 | 25 | def __init__(self, connect_url, worker_id, unhealthy_states, auth, failure_threshold_percentage, considered_containers): 26 | self.connect_url = connect_url 27 | self.worker_id = worker_id 28 | self.unhealthy_states = [x.upper().strip() for x in unhealthy_states] 29 | self.failure_threshold = failure_threshold_percentage * .01 30 | self.considered_containers = [x.lower().strip() for x in considered_containers] 31 | self.kwargs = {} 32 | if auth and ":" in auth: 33 | self.kwargs["auth"] = tuple(auth.split(":")) 34 | self.log_initialization_values() 35 | 36 | def get_health_result(self): 37 | try: 38 | health_result = {"failures": [], "failure_states": self.unhealthy_states} 39 | connector_names = self.get_connector_names() 40 | connector_statuses = self.get_connectors_health(connector_names) 41 | self.handle_healthcheck(connector_statuses, health_result) 42 | 43 | connector_count = len(connector_names) 44 | task_count = sum(len(c["tasks"]) for c in connector_statuses) 45 | 46 | container_count = 0 47 | if "connector" in self.considered_containers: 48 | container_count += connector_count 49 | if "task" in self.considered_containers: 50 | container_count += task_count 51 | 52 | failure_count = len([f for f in health_result["failures"] if f["type"] in self.considered_containers]) 53 | 54 | # guards against division by zero. if we have no connectors or tasks we are deciding to pass 55 | if container_count > 0: 56 | health_result["failure_rate"] = failure_count / container_count 57 | else: 58 | health_result["failure_rate"] = 0.0 59 | 60 | health_result["failure_threshold"] = self.failure_threshold 61 | health_result["healthy"] = health_result["failure_rate"] <= health_result["failure_threshold"] 62 | 63 | # broker errors override any failure calculation 64 | if any([f for f in health_result["failures"] if f["type"] == "broker"]): 65 | health_result["healthy"] = False 66 | 67 | except Exception as ex: 68 | logging.error("Error while attempting to calculate health result. Assuming unhealthy. Error: {}".format(ex)) 69 | logging.error(ex) 70 | health_result = { 71 | "healthy": False, 72 | "message": "Exception raised while attempting to calculate health result, assuming unhealthy.", 73 | "error": "{}".format(ex), 74 | "failure_states": self.unhealthy_states 75 | } 76 | helpers.log_line_break() 77 | return health_result 78 | 79 | def handle_healthcheck(self, connector_statuses, health_result): 80 | connectors_on_this_worker = False 81 | for connector in connector_statuses: 82 | if self.is_on_this_worker(connector["worker_id"]) and "connector" in self.considered_containers: 83 | connectors_on_this_worker = True 84 | if self.is_in_unhealthy_state(connector["state"]): 85 | logging.warning("Connector '{}' is unhealthy in failure state: {}".format(connector["name"], connector["state"])) 86 | health_result["failures"].append({ 87 | "type": "connector", 88 | "connector": connector["name"], 89 | "state": connector["state"], 90 | "worker_id": connector["worker_id"] 91 | }) 92 | else: 93 | logging.info("Connector '{}' is healthy in state: {}".format(connector["name"], connector["state"])) 94 | self.handle_task_healthcheck(connector, health_result) 95 | if not connectors_on_this_worker and connector_statuses: 96 | self.handle_broker_healthcheck(health_result, connector_statuses[0]["name"]) 97 | 98 | def handle_broker_healthcheck(self, health_result, connector_name): 99 | try: 100 | self.get_connector_details(connector_name) 101 | except Exception as ex: 102 | logging.error("Error while attempting to get details for {}. Assuming unhealthy. Error: {}".format(connector_name, ex)) 103 | logging.error(ex) 104 | health_result["failures"].append({ 105 | "type": "broker", 106 | "connector": connector_name, 107 | }) 108 | 109 | def handle_task_healthcheck(self, connector, health_result): 110 | if "task" in self.considered_containers: 111 | for task in connector["tasks"]: 112 | if self.is_on_this_worker(task["worker_id"]): 113 | if self.is_in_unhealthy_state(task["state"]): 114 | logging.warning("Connector '{}' task '{}' is unhealthy in failure state: {}".format( 115 | connector["name"], task["id"], task["state"] 116 | )) 117 | health_result["failures"].append({ 118 | "type": "task", 119 | "connector": connector["name"], 120 | "id": task["id"], 121 | "state": task["state"], 122 | "worker_id": task["worker_id"], 123 | "trace": task.get("trace", None) 124 | }) 125 | else: 126 | logging.info("Connector '{}' task '{}' is healthy in state: {}".format( 127 | connector["name"], task["id"], task["state"] 128 | )) 129 | 130 | def get_connectors_health(self, connector_names): 131 | statuses = [] 132 | for connector_name in connector_names: 133 | statuses.append(self.get_connector_health(connector_name)) 134 | return statuses 135 | 136 | def get_connector_health(self, connector_name): 137 | connector_status = self.get_connector_status(connector_name) 138 | connector_state = connector_status["connector"]["state"].upper() 139 | connector_worker = connector_status["connector"]["worker_id"] 140 | return { 141 | "name": connector_name, 142 | "state": connector_state, 143 | "worker_id": connector_worker, 144 | "tasks": connector_status["tasks"] 145 | } 146 | 147 | def get_connector_names(self): 148 | response = requests.get("{}/connectors".format(self.connect_url), **self.kwargs) 149 | response_json = response.json() 150 | return response_json 151 | 152 | def get_connector_status(self, connector_name): 153 | response = requests.get("{}/connectors/{}/status".format(self.connect_url, connector_name), **self.kwargs) 154 | response_json = response.json() 155 | return response_json 156 | 157 | def get_connector_details(self, connector_name): 158 | response = requests.get("{}/connectors/{}".format(self.connect_url, connector_name), **self.kwargs) 159 | response.raise_for_status() 160 | response_json = response.json() 161 | return response_json 162 | 163 | def is_in_unhealthy_state(self, state): 164 | return state.upper() in self.unhealthy_states 165 | 166 | def is_on_this_worker(self, response_worker_id): 167 | return response_worker_id.lower() == self.worker_id.lower() if self.worker_id is not None else True 168 | 169 | def log_initialization_values(self): 170 | logging.info("Server will report unhealthy for states: '{}'".format(", ".join(self.unhealthy_states))) 171 | logging.info("Server will healthcheck against Kafka Connect at: {}".format(self.connect_url)) 172 | if "auth" in self.kwargs: 173 | logging.info("Server will use basic authentication against Kafka Connect") 174 | if self.worker_id is not None: 175 | logging.info("Server will healthcheck connectors and tasks for worker with id '{}'".format(self.worker_id)) 176 | else: 177 | logging.warning("No worker id supplied, server will healthcheck all connectors and tasks") 178 | -------------------------------------------------------------------------------- /kafka_connect_healthcheck/helpers.py: -------------------------------------------------------------------------------- 1 | # -*- coding: utf-8 -*- 2 | # 3 | # Copyright 2019 Shawn Seymour. All Rights Reserved. 4 | # 5 | # Licensed under the Apache License, Version 2.0 (the "License"). You 6 | # may not use this file except in compliance with the License. A copy of 7 | # the License is located at 8 | # 9 | # http://www.apache.org/licenses/LICENSE-2.0 10 | # 11 | # or in the "license" file accompanying this file. This file is 12 | # distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF 13 | # ANY KIND, either express or implied. See the License for the specific 14 | # language governing permissions and limitations under the License. 15 | 16 | import logging 17 | 18 | 19 | def log_line_break(): 20 | logging.info("------------------------------------------------") 21 | -------------------------------------------------------------------------------- /kafka_connect_healthcheck/main.py: -------------------------------------------------------------------------------- 1 | # -*- coding: utf-8 -*- 2 | # 3 | # Copyright 2019 Shawn Seymour. All Rights Reserved. 4 | # 5 | # Licensed under the Apache License, Version 2.0 (the "License"). You 6 | # may not use this file except in compliance with the License. A copy of 7 | # the License is located at 8 | # 9 | # http://www.apache.org/licenses/LICENSE-2.0 10 | # 11 | # or in the "license" file accompanying this file. This file is 12 | # distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF 13 | # ANY KIND, either express or implied. See the License for the specific 14 | # language governing permissions and limitations under the License. 15 | 16 | import logging 17 | import signal 18 | import sys 19 | from functools import partial 20 | from http.server import HTTPServer 21 | 22 | from kafka_connect_healthcheck import health 23 | from kafka_connect_healthcheck import helpers 24 | from kafka_connect_healthcheck import parser 25 | from kafka_connect_healthcheck.handler import RequestHandler 26 | 27 | 28 | def main(): 29 | config_parser = parser.get_parser() 30 | args = config_parser.parse_args() 31 | 32 | logging.basicConfig(format="%(asctime)s.%(msecs)03d [%(levelname)7s] - %(message)s", 33 | datefmt="%Y-%m-%d %H:%M:%S", 34 | level=args.log_level) 35 | 36 | logging.info("Initializing healthcheck server...") 37 | 38 | server_class = HTTPServer 39 | health_object = health.Health(args.connect_url, args.connect_worker_id, args.unhealthy_states.split(","), 40 | args.basic_auth, args.failure_threshold_percentage, args.considered_containers.split(",")) 41 | handler = partial(RequestHandler, health_object) 42 | httpd = server_class(("0.0.0.0", args.healthcheck_port), handler) 43 | logging.info("Healthcheck server started at: http://localhost:{}".format(args.healthcheck_port)) 44 | helpers.log_line_break() 45 | 46 | def stop(status_code, frame): 47 | logging.info("SIGINT/SIGTERM; exiting...") 48 | httpd.server_close() 49 | sys.exit(0) 50 | 51 | signal.signal(signal.SIGINT, stop) 52 | signal.signal(signal.SIGTERM, stop) 53 | 54 | try: 55 | httpd.serve_forever() 56 | except KeyboardInterrupt: 57 | stop(0, None) 58 | 59 | 60 | if __name__ == "__main__": 61 | main() 62 | -------------------------------------------------------------------------------- /kafka_connect_healthcheck/parser.py: -------------------------------------------------------------------------------- 1 | # -*- coding: utf-8 -*- 2 | # 3 | # Copyright 2019 Shawn Seymour. All Rights Reserved. 4 | # 5 | # Licensed under the Apache License, Version 2.0 (the "License"). You 6 | # may not use this file except in compliance with the License. A copy of 7 | # the License is located at 8 | # 9 | # http://www.apache.org/licenses/LICENSE-2.0 10 | # 11 | # or in the "license" file accompanying this file. This file is 12 | # distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF 13 | # ANY KIND, either express or implied. See the License for the specific 14 | # language governing permissions and limitations under the License. 15 | 16 | import argparse 17 | import os 18 | 19 | 20 | def get_parser(): 21 | parser = argparse.ArgumentParser(description="A simple healthcheck for Kafka Connect. It can wrap a Kafka Connect " 22 | "process so you can check the health of specific connectors. " 23 | "This can be used as a health check within Kubernetes to automatically " 24 | "restart failed connect instances.") 25 | 26 | parser.add_argument("--port", 27 | default=os.environ.get("HEALTHCHECK_PORT", "18083"), 28 | dest="healthcheck_port", 29 | type=int, 30 | nargs="?", 31 | help="The port for the healthcheck HTTP server." 32 | ) 33 | 34 | parser.add_argument("--connect-url", 35 | default=os.environ.get("HEALTHCHECK_CONNECT_URL", "http://localhost:8083"), 36 | dest="connect_url", 37 | nargs="?", 38 | help="The Kafka Connect REST API URL that the health check will be run against." 39 | ) 40 | 41 | parser.add_argument("--connect-worker-id", 42 | default=os.environ.get("HEALTHCHECK_CONNECT_WORKER_ID"), 43 | dest="connect_worker_id", 44 | nargs="?", 45 | help="The Kafka Connect REST API URL that the health check will be run against." 46 | ) 47 | 48 | parser.add_argument("--unhealthy-states", 49 | default=os.environ.get("HEALTHCHECK_UNHEALTHY_STATES", "FAILED").upper(), 50 | dest="unhealthy_states", 51 | nargs="?", 52 | help="A comma separated lists of connector and task states to be marked as unhealthy. Default: FAILED." 53 | ) 54 | 55 | parser.add_argument("--considered-containers", 56 | default=os.environ.get("HEALTHCHECK_CONSIDERED_CONTAINERS", "CONNECTOR,TASK").upper(), 57 | dest="considered_containers", 58 | nargs="?", 59 | help="A comma separated lists of container types to consider for failure calculations. Default: CONNECTOR,TASK." 60 | ) 61 | 62 | parser.add_argument("--failure-threshold-percentage", 63 | default=os.environ.get("HEALTHCHECK_FAILURE_THRESHOLD_PERCENTAGE", 0), 64 | dest="failure_threshold_percentage", 65 | type=int, 66 | nargs="?", 67 | help="A number between 1 and 100. If set, this is the percentage of connectors that must fail for the healthcheck to fail." 68 | ) 69 | 70 | parser.add_argument("--basic-auth", 71 | default=os.environ.get("HEALTHCHECK_BASIC_AUTH", ""), 72 | dest="basic_auth", 73 | nargs="?", 74 | help="Colon-separated credentials for basic HTTP authentication. Default: empty.") 75 | 76 | parser.add_argument("--log-level", 77 | default=os.environ.get("HEALTHCHECK_LOG_LEVEL", "INFO").upper(), 78 | dest="log_level", 79 | nargs="?", 80 | help="The level of logs to be shown. Default: INFO.") 81 | 82 | return parser 83 | -------------------------------------------------------------------------------- /kafka_connect_healthcheck/version.py: -------------------------------------------------------------------------------- 1 | # -*- coding: utf-8 -*- 2 | # 3 | # Copyright 2019 Shawn Seymour. All Rights Reserved. 4 | # 5 | # Licensed under the Apache License, Version 2.0 (the "License"). You 6 | # may not use this file except in compliance with the License. A copy of 7 | # the License is located at 8 | # 9 | # http://www.apache.org/licenses/LICENSE-2.0 10 | # 11 | # or in the "license" file accompanying this file. This file is 12 | # distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF 13 | # ANY KIND, either express or implied. See the License for the specific 14 | # language governing permissions and limitations under the License. 15 | 16 | __version__ = "0.3.0" 17 | 18 | 19 | def get_version(): 20 | return __version__ 21 | -------------------------------------------------------------------------------- /requirements-dev.txt: -------------------------------------------------------------------------------- 1 | pytest>=2.7.0 2 | pytest-cov>=2.4.0,<2.6 3 | mock>=1.0.1 4 | tox>=1.9.2 5 | flake8>=3.7.7 6 | flake8-quotes>=1.0.0 7 | coverage>=4.5.2 8 | requests>=2.21.0 9 | -------------------------------------------------------------------------------- /scripts/build-docker.sh: -------------------------------------------------------------------------------- 1 | #!/usr/bin/env bash 2 | 3 | docker build -t devshawn/kafka-connect-healthcheck . 4 | -------------------------------------------------------------------------------- /scripts/lint.sh: -------------------------------------------------------------------------------- 1 | #!/usr/bin/env bash 2 | 3 | flake8 --max-line-length=135 --inline-quotes '"' --exclude=.tox,htmlcov,build,tests,scratch,docs,venv . 4 | -------------------------------------------------------------------------------- /scripts/test-coverage.sh: -------------------------------------------------------------------------------- 1 | #!/usr/bin/env bash 2 | 3 | python3 -m pytest --cov=kafka_connect_healthcheck -vs --cov-report html -------------------------------------------------------------------------------- /scripts/test.sh: -------------------------------------------------------------------------------- 1 | #!/usr/bin/env bash 2 | 3 | python3 -m pytest -------------------------------------------------------------------------------- /setup.cfg: -------------------------------------------------------------------------------- 1 | [metadata] 2 | description-file = README.md -------------------------------------------------------------------------------- /setup.py: -------------------------------------------------------------------------------- 1 | import setuptools 2 | 3 | meta = {} 4 | 5 | with open("README.md", "r") as fh: 6 | long_description = fh.read() 7 | 8 | with open("kafka_connect_healthcheck/version.py") as f: 9 | exec(f.read(), meta) 10 | 11 | requires = [ 12 | "requests>=2.21.0" 13 | ] 14 | 15 | setuptools.setup( 16 | name="kafka-connect-healthcheck", 17 | version=meta["__version__"], 18 | author="Shawn Seymour", 19 | author_email="shawn@devshawn.com", 20 | description="A simple healthcheck wrapper to monitor Kafka Connect.", 21 | long_description=long_description, 22 | long_description_content_type="text/markdown", 23 | url="https://github.com/devshawn/kafka-connect-healthcheck", 24 | license="Apache License 2.0", 25 | packages=["kafka_connect_healthcheck"], 26 | install_requires=requires, 27 | entry_points={ 28 | "console_scripts": ["kafka-connect-healthcheck=kafka_connect_healthcheck.main:main"], 29 | }, 30 | keywords=("kafka", "connect", "health", "healthcheck", "wrapper", "monitor", "connector"), 31 | classifiers=[ 32 | "Development Status :: 4 - Beta", 33 | "Intended Audience :: Developers", 34 | "Intended Audience :: System Administrators", 35 | "License :: OSI Approved :: Apache Software License", 36 | "Natural Language :: English", 37 | "Operating System :: MacOS", 38 | "Operating System :: Unix", 39 | "Programming Language :: Python :: 3", 40 | "Programming Language :: Python :: 3.5", 41 | "Programming Language :: Python :: 3.6", 42 | "Programming Language :: Python :: 3.7", 43 | "Programming Language :: Python :: 3.8", 44 | "Topic :: Software Development", 45 | "Topic :: Software Development :: Libraries :: Python Modules", 46 | ], 47 | ) 48 | -------------------------------------------------------------------------------- /tests/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/devshawn/kafka-connect-healthcheck/03645ba3cf877de69f615a360441b51c6319ffb6/tests/__init__.py -------------------------------------------------------------------------------- /tests/conftest.py: -------------------------------------------------------------------------------- 1 | import os 2 | import signal 3 | import subprocess 4 | import time 5 | 6 | import coverage.data 7 | import pytest 8 | import requests 9 | from requests.exceptions import ConnectionError 10 | 11 | from tests import mocks 12 | 13 | 14 | @pytest.fixture(autouse=True) 15 | def run_backend(cov, request): 16 | env = os.environ.copy() 17 | env['COVERAGE_FILE'] = '.coverage.backend' 18 | 19 | try: 20 | test_environment, mock_name = request.param 21 | kafka_connect_url = None 22 | if test_environment.get("HEALTHCHECK_CONNECT_URL") is None: 23 | kafka_connect_port = setup_mock_kafka_connect_api(mock_name) 24 | kafka_connect_url = "http://localhost:{}".format(kafka_connect_port) 25 | env['HEALTHCHECK_CONNECT_URL'] = kafka_connect_url 26 | 27 | for key, value in test_environment.items(): 28 | env[key] = value 29 | print("\nAdding environment variable: {} --> {}".format(key, value)) 30 | 31 | process = generate_subprocess(env, request) 32 | wait_for_servers_to_be_up(kafka_connect_url) 33 | 34 | yield 35 | 36 | process.send_signal(signal.SIGINT) 37 | # out, err = process.communicate() 38 | # request.config.cache.set('logs', err.decode("utf-8")) 39 | time.sleep(1) 40 | 41 | if is_coverage_on(request): 42 | write_backend_coverage(cov) 43 | 44 | except AttributeError: 45 | print("\nNo parameters, ignoring fixture.") 46 | yield 47 | 48 | 49 | def generate_subprocess(env, request): 50 | print() 51 | if is_coverage_on(request): 52 | process_args = " ".join(['exec', 'python3', '-m', 'coverage', 'run', '--source', './kafka_connect_healthcheck', 53 | './kafka_connect_healthcheck/main.py']), 54 | else: 55 | process_args = " ".join(['exec', 'python3', './kafka_connect_healthcheck/main.py']), 56 | return subprocess.Popen( 57 | process_args, 58 | env=env, 59 | shell=True, 60 | stdout=None if os.environ.get("TEST_SHOW_SERVER_LOGS", "false") == "true" else subprocess.PIPE, 61 | stderr=None if os.environ.get("TEST_SHOW_SERVER_LOGS", "false") == "true" else subprocess.PIPE, 62 | preexec_fn=os.setsid 63 | ) 64 | 65 | 66 | def write_backend_coverage(cov): 67 | backendcov = coverage.data.CoverageData() 68 | with open('.coverage.backend') as fp: 69 | backendcov.read_fileobj(fp) 70 | cov.data.update(backendcov) 71 | 72 | 73 | def setup_mock_kafka_connect_api(mock_name): 74 | kafka_connect_port = mocks.get_free_port() 75 | mocks.start_mock_server(kafka_connect_port, mock_name) 76 | return kafka_connect_port 77 | 78 | 79 | def wait_for_servers_to_be_up(kafka_connect_url): 80 | is_alive = False 81 | while is_alive is False: 82 | is_alive = are_servers_up("http://localhost:18083", kafka_connect_url) 83 | time.sleep(0.2) 84 | 85 | 86 | def are_servers_up(healthcheck_url, kafka_connect_url): 87 | try: 88 | if kafka_connect_url is not None: 89 | response = requests.head(healthcheck_url) 90 | is_mock_server_alive = response.status_code == 200 91 | else: 92 | is_mock_server_alive = True 93 | response = requests.head(healthcheck_url) 94 | return response.status_code == 200 and is_mock_server_alive 95 | except ConnectionError: 96 | return False 97 | 98 | 99 | def is_coverage_on(request): 100 | return len(request.config.getoption('--cov')) != 0 101 | -------------------------------------------------------------------------------- /tests/context.py: -------------------------------------------------------------------------------- 1 | import sys 2 | import os 3 | 4 | sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), '../kafka_connect_healthcheck'))) 5 | 6 | import kafka_connect_healthcheck 7 | -------------------------------------------------------------------------------- /tests/data/expected/0-healthcheck-server-healthy.json: -------------------------------------------------------------------------------- 1 | { 2 | "status": "UP" 3 | } -------------------------------------------------------------------------------- /tests/data/expected/1-healthy.json: -------------------------------------------------------------------------------- 1 | { 2 | "failures": [], 3 | "failure_states": [ 4 | "FAILED" 5 | ], 6 | "failure_rate": 0.0, 7 | "failure_threshold": 0.0, 8 | "healthy": true 9 | } 10 | -------------------------------------------------------------------------------- /tests/data/expected/10-unhealthy-multiple-connectors.json: -------------------------------------------------------------------------------- 1 | { 2 | "failures": [ 3 | { 4 | "type": "connector", 5 | "connector": "s3-source", 6 | "state": "FAILED", 7 | "worker_id": "my.worker.name:8083" 8 | }, 9 | { 10 | "type": "task", 11 | "connector": "s3-source", 12 | "id": 3, 13 | "state": "FAILED", 14 | "worker_id": "my.worker.name:8083", 15 | "trace": null 16 | }, 17 | { 18 | "type": "task", 19 | "connector": "s3-sink", 20 | "id": 2, 21 | "state": "FAILED", 22 | "worker_id": "my.worker.name:8083", 23 | "trace": null 24 | } 25 | ], 26 | "failure_states": [ 27 | "FAILED" 28 | ], 29 | "failure_rate": 0.3, 30 | "failure_threshold": 0.0, 31 | "healthy": false 32 | } 33 | -------------------------------------------------------------------------------- /tests/data/expected/11-healthy-no-connectors.json: -------------------------------------------------------------------------------- 1 | { 2 | "failures": [], 3 | "failure_states": [ 4 | "FAILED" 5 | ], 6 | "failure_rate": 0.0, 7 | "failure_threshold": 0.0, 8 | "healthy": true 9 | } 10 | -------------------------------------------------------------------------------- /tests/data/expected/12-unhealthy-task-with-trace.json: -------------------------------------------------------------------------------- 1 | { 2 | "failures": [ 3 | { 4 | "type": "task", 5 | "connector": "s3-sink", 6 | "id": 2, 7 | "state": "FAILED", 8 | "worker_id": "my.worker.name:8083", 9 | "trace": "trace message here" 10 | } 11 | ], 12 | "failure_states": [ 13 | "FAILED" 14 | ], 15 | "failure_rate": 0.3333333333333333, 16 | "failure_threshold": 0.0, 17 | "healthy": false 18 | } 19 | -------------------------------------------------------------------------------- /tests/data/expected/13-unhealthy-broker-connection.json: -------------------------------------------------------------------------------- 1 | { 2 | "failures": [ 3 | { 4 | "type": "broker", 5 | "connector": "s3-sink" 6 | } 7 | ], 8 | "failure_states": [ 9 | "FAILED" 10 | ], 11 | "failure_rate": 0.0, 12 | "failure_threshold": 0.0, 13 | "healthy": false 14 | } 15 | -------------------------------------------------------------------------------- /tests/data/expected/14-basic-auth.json: -------------------------------------------------------------------------------- 1 | { 2 | "failures": [], 3 | "failure_states": [ 4 | "FAILED" 5 | ], 6 | "failure_rate": 0.0, 7 | "failure_threshold": 0.0, 8 | "healthy": true 9 | } 10 | -------------------------------------------------------------------------------- /tests/data/expected/15-unhealthy-threshold.json: -------------------------------------------------------------------------------- 1 | { 2 | "failures": [ 3 | { 4 | "type": "connector", 5 | "connector": "s3-source", 6 | "state": "FAILED", 7 | "worker_id": "my.worker.name:8083" 8 | }, 9 | { 10 | "type": "task", 11 | "connector": "s3-source", 12 | "id": 3, 13 | "state": "FAILED", 14 | "worker_id": "my.worker.name:8083", 15 | "trace": null 16 | }, 17 | { 18 | "type": "task", 19 | "connector": "s3-sink", 20 | "id": 2, 21 | "state": "FAILED", 22 | "worker_id": "my.worker.name:8083", 23 | "trace": null 24 | } 25 | ], 26 | "failure_states": [ 27 | "FAILED" 28 | ], 29 | "failure_rate": 0.3, 30 | "failure_threshold": 0.1, 31 | "healthy": false 32 | } 33 | -------------------------------------------------------------------------------- /tests/data/expected/16-healthy-threshold.json: -------------------------------------------------------------------------------- 1 | { 2 | "failures": [ 3 | { 4 | "type": "connector", 5 | "connector": "s3-source", 6 | "state": "FAILED", 7 | "worker_id": "my.worker.name:8083" 8 | }, 9 | { 10 | "type": "task", 11 | "connector": "s3-source", 12 | "id": 3, 13 | "state": "FAILED", 14 | "worker_id": "my.worker.name:8083", 15 | "trace": null 16 | }, 17 | { 18 | "type": "task", 19 | "connector": "s3-sink", 20 | "id": 2, 21 | "state": "FAILED", 22 | "worker_id": "my.worker.name:8083", 23 | "trace": null 24 | } 25 | ], 26 | "failure_states": [ 27 | "FAILED" 28 | ], 29 | "failure_rate": 0.3, 30 | "failure_threshold": 0.5, 31 | "healthy": true 32 | } 33 | -------------------------------------------------------------------------------- /tests/data/expected/17-healthy-container-connector.json: -------------------------------------------------------------------------------- 1 | { 2 | "failures": [], 3 | "failure_states": [ 4 | "FAILED" 5 | ], 6 | "failure_rate": 0.0, 7 | "failure_threshold": 0.0, 8 | "healthy": true 9 | } 10 | -------------------------------------------------------------------------------- /tests/data/expected/18-healthy-container-task.json: -------------------------------------------------------------------------------- 1 | { 2 | "failures": [], 3 | "failure_states": [ 4 | "FAILED" 5 | ], 6 | "failure_rate": 0.0, 7 | "failure_threshold": 0.0, 8 | "healthy": true 9 | } 10 | -------------------------------------------------------------------------------- /tests/data/expected/19-unhealthy-container-task.json: -------------------------------------------------------------------------------- 1 | { 2 | "failures": [ 3 | { 4 | "type": "task", 5 | "connector": "local-file-source", 6 | "id": 0, 7 | "state": "FAILED", 8 | "worker_id": "kafka-connect:8083", 9 | "trace": null 10 | } 11 | ], 12 | "failure_states": [ 13 | "FAILED" 14 | ], 15 | "failure_rate": 1.0, 16 | "failure_threshold": 0.0, 17 | "healthy": false 18 | } 19 | -------------------------------------------------------------------------------- /tests/data/expected/2-unhealthy.json: -------------------------------------------------------------------------------- 1 | { 2 | "failures": [ 3 | { 4 | "type": "connector", 5 | "connector": "local-file-source", 6 | "state": "FAILED", 7 | "worker_id": "kafka-connect:8083" 8 | }, 9 | { 10 | "type": "task", 11 | "connector": "local-file-source", 12 | "id": 0, 13 | "state": "FAILED", 14 | "worker_id": "kafka-connect:8083", 15 | "trace": null 16 | } 17 | ], 18 | "failure_states": [ 19 | "FAILED" 20 | ], 21 | "failure_rate": 1.0, 22 | "failure_threshold": 0.0, 23 | "healthy": false 24 | } 25 | -------------------------------------------------------------------------------- /tests/data/expected/3-unhealthy-error.json: -------------------------------------------------------------------------------- 1 | { 2 | "healthy": false, 3 | "message": "Exception raised while attempting to calculate health result, assuming unhealthy.", 4 | "error": "HTTPConnectionPool(host='unknown-hostname', port=8083): Max retries exceeded with url: /connectors", 5 | "failure_states": [ 6 | "FAILED" 7 | ] 8 | } 9 | -------------------------------------------------------------------------------- /tests/data/expected/4-healthy-worker-id-correct.json: -------------------------------------------------------------------------------- 1 | { 2 | "failures": [], 3 | "failure_states": [ 4 | "FAILED" 5 | ], 6 | "failure_rate": 0.0, 7 | "failure_threshold": 0.0, 8 | "healthy": true 9 | } 10 | -------------------------------------------------------------------------------- /tests/data/expected/5-healthy-worker-id-unused.json: -------------------------------------------------------------------------------- 1 | { 2 | "failures": [], 3 | "failure_states": [ 4 | "FAILED" 5 | ], 6 | "failure_rate": 0.0, 7 | "failure_threshold": 0.0, 8 | "healthy": true 9 | } 10 | -------------------------------------------------------------------------------- /tests/data/expected/6-healthy-worker-id-with-other-workers-failing.json: -------------------------------------------------------------------------------- 1 | { 2 | "failures": [], 3 | "failure_states": [ 4 | "FAILED" 5 | ], 6 | "failure_rate": 0.0, 7 | "failure_threshold": 0.0, 8 | "healthy": true 9 | } 10 | -------------------------------------------------------------------------------- /tests/data/expected/7-unhealthy-worker-id-with-other-workers-healthy.json: -------------------------------------------------------------------------------- 1 | { 2 | "failures": [ 3 | { 4 | "type": "task", 5 | "connector": "jdbc-sink", 6 | "id": 0, 7 | "state": "FAILED", 8 | "worker_id": "kafka-connect:8083", 9 | "trace": null 10 | } 11 | ], 12 | "failure_states": [ 13 | "FAILED" 14 | ], 15 | "failure_rate": 0.5, 16 | "failure_threshold": 0.0, 17 | "healthy": false 18 | } 19 | -------------------------------------------------------------------------------- /tests/data/expected/8-healthy-multiple-tasks.json: -------------------------------------------------------------------------------- 1 | { 2 | "failures": [], 3 | "failure_states": [ 4 | "FAILED" 5 | ], 6 | "failure_rate": 0.0, 7 | "failure_threshold": 0.0, 8 | "healthy": true 9 | } 10 | -------------------------------------------------------------------------------- /tests/data/expected/9-healthy-multiple-connectors.json: -------------------------------------------------------------------------------- 1 | { 2 | "failures": [], 3 | "failure_states": [ 4 | "FAILED" 5 | ], 6 | "failure_rate": 0.0, 7 | "failure_threshold": 0.0, 8 | "healthy": true 9 | } 10 | -------------------------------------------------------------------------------- /tests/data/mocks/1-healthy-connector-local-file-source.json: -------------------------------------------------------------------------------- 1 | { 2 | "name": "local-file-source", 3 | "connector": { 4 | "state": "RUNNING", 5 | "worker_id": "kafka-connect:8083" 6 | }, 7 | "tasks": [ 8 | { 9 | "id": 0, 10 | "state": "RUNNING", 11 | "worker_id": "kafka-connect:8083" 12 | } 13 | ], 14 | "type": "source" 15 | } -------------------------------------------------------------------------------- /tests/data/mocks/1-healthy-connectors.json: -------------------------------------------------------------------------------- 1 | [ 2 | "local-file-source" 3 | ] -------------------------------------------------------------------------------- /tests/data/mocks/10-unhealthy-multiple-connectors-connector-s3-sink.json: -------------------------------------------------------------------------------- 1 | { 2 | "name": "s3-sink", 3 | "connector": { 4 | "state": "RUNNING", 5 | "worker_id": "my.worker.name:8083" 6 | }, 7 | "tasks": [ 8 | { 9 | "id": 0, 10 | "state": "RUNNING", 11 | "worker_id": "my.worker.name:8083" 12 | }, 13 | { 14 | "id": 2, 15 | "state": "FAILED", 16 | "worker_id": "my.worker.name:8083" 17 | }, 18 | { 19 | "id": 3, 20 | "state": "RUNNING", 21 | "worker_id": "my.worker.name:8083" 22 | }, 23 | { 24 | "id": 4, 25 | "state": "RUNNING", 26 | "worker_id": "my.worker.name:8083" 27 | } 28 | ], 29 | "type": "source" 30 | } -------------------------------------------------------------------------------- /tests/data/mocks/10-unhealthy-multiple-connectors-connector-s3-source.json: -------------------------------------------------------------------------------- 1 | { 2 | "name": "s3-source", 3 | "connector": { 4 | "state": "FAILED", 5 | "worker_id": "my.worker.name:8083" 6 | }, 7 | "tasks": [ 8 | { 9 | "id": 0, 10 | "state": "RUNNING", 11 | "worker_id": "my.worker.name:8083" 12 | }, 13 | { 14 | "id": 2, 15 | "state": "RUNNING", 16 | "worker_id": "my.worker.name:8083" 17 | }, 18 | { 19 | "id": 3, 20 | "state": "FAILED", 21 | "worker_id": "my.worker.name:8083" 22 | }, 23 | { 24 | "id": 4, 25 | "state": "RUNNING", 26 | "worker_id": "my.worker.name:8083" 27 | } 28 | ], 29 | "type": "source" 30 | } -------------------------------------------------------------------------------- /tests/data/mocks/10-unhealthy-multiple-connectors-connectors.json: -------------------------------------------------------------------------------- 1 | [ 2 | "s3-source", 3 | "s3-sink" 4 | ] -------------------------------------------------------------------------------- /tests/data/mocks/11-healthy-no-connectors-connectors.json: -------------------------------------------------------------------------------- 1 | [] 2 | -------------------------------------------------------------------------------- /tests/data/mocks/12-unhealthy-task-with-trace-connector-s3-sink.json: -------------------------------------------------------------------------------- 1 | { 2 | "name": "s3-sink", 3 | "connector": { 4 | "state": "RUNNING", 5 | "worker_id": "my.worker.name:8083" 6 | }, 7 | "tasks": [ 8 | { 9 | "id": 0, 10 | "state": "RUNNING", 11 | "worker_id": "my.worker.name:8083" 12 | }, 13 | { 14 | "id": 2, 15 | "state": "FAILED", 16 | "worker_id": "my.worker.name:8083", 17 | "trace": "trace message here" 18 | } 19 | ], 20 | "type": "source" 21 | } 22 | -------------------------------------------------------------------------------- /tests/data/mocks/12-unhealthy-task-with-trace-connectors.json: -------------------------------------------------------------------------------- 1 | [ 2 | "s3-sink" 3 | ] 4 | -------------------------------------------------------------------------------- /tests/data/mocks/13-unhealthy-broker-connection-connector-s3-sink.json: -------------------------------------------------------------------------------- 1 | { 2 | "name": "s3-sink", 3 | "connector": { 4 | "state": "RUNNING", 5 | "worker_id": "my.worker.name:8083" 6 | }, 7 | "tasks": [ 8 | { 9 | "id": 0, 10 | "state": "RUNNING", 11 | "worker_id": "other.worker:8083" 12 | } 13 | ], 14 | "type": "source" 15 | } 16 | -------------------------------------------------------------------------------- /tests/data/mocks/13-unhealthy-broker-connection-connectors.json: -------------------------------------------------------------------------------- 1 | [ 2 | "s3-sink" 3 | ] 4 | -------------------------------------------------------------------------------- /tests/data/mocks/14-basic-auth-connector-local-file-source.json: -------------------------------------------------------------------------------- 1 | { 2 | "name": "local-file-source", 3 | "connector": { 4 | "state": "RUNNING", 5 | "worker_id": "kafka-connect:8083" 6 | }, 7 | "tasks": [ 8 | { 9 | "id": 0, 10 | "state": "RUNNING", 11 | "worker_id": "kafka-connect:8083" 12 | } 13 | ], 14 | "type": "source" 15 | } 16 | -------------------------------------------------------------------------------- /tests/data/mocks/14-basic-auth-connectors.json: -------------------------------------------------------------------------------- 1 | [ 2 | "local-file-source" 3 | ] 4 | -------------------------------------------------------------------------------- /tests/data/mocks/15-unhealthy-threshold-connector-s3-sink.json: -------------------------------------------------------------------------------- 1 | { 2 | "name": "s3-sink", 3 | "connector": { 4 | "state": "RUNNING", 5 | "worker_id": "my.worker.name:8083" 6 | }, 7 | "tasks": [ 8 | { 9 | "id": 0, 10 | "state": "RUNNING", 11 | "worker_id": "my.worker.name:8083" 12 | }, 13 | { 14 | "id": 2, 15 | "state": "FAILED", 16 | "worker_id": "my.worker.name:8083" 17 | }, 18 | { 19 | "id": 3, 20 | "state": "RUNNING", 21 | "worker_id": "my.worker.name:8083" 22 | }, 23 | { 24 | "id": 4, 25 | "state": "RUNNING", 26 | "worker_id": "my.worker.name:8083" 27 | } 28 | ], 29 | "type": "source" 30 | } -------------------------------------------------------------------------------- /tests/data/mocks/15-unhealthy-threshold-connector-s3-source.json: -------------------------------------------------------------------------------- 1 | { 2 | "name": "s3-source", 3 | "connector": { 4 | "state": "FAILED", 5 | "worker_id": "my.worker.name:8083" 6 | }, 7 | "tasks": [ 8 | { 9 | "id": 0, 10 | "state": "RUNNING", 11 | "worker_id": "my.worker.name:8083" 12 | }, 13 | { 14 | "id": 2, 15 | "state": "RUNNING", 16 | "worker_id": "my.worker.name:8083" 17 | }, 18 | { 19 | "id": 3, 20 | "state": "FAILED", 21 | "worker_id": "my.worker.name:8083" 22 | }, 23 | { 24 | "id": 4, 25 | "state": "RUNNING", 26 | "worker_id": "my.worker.name:8083" 27 | } 28 | ], 29 | "type": "source" 30 | } -------------------------------------------------------------------------------- /tests/data/mocks/15-unhealthy-threshold-connectors.json: -------------------------------------------------------------------------------- 1 | [ 2 | "s3-source", 3 | "s3-sink" 4 | ] -------------------------------------------------------------------------------- /tests/data/mocks/16-healthy-threshold-connector-s3-sink.json: -------------------------------------------------------------------------------- 1 | { 2 | "name": "s3-sink", 3 | "connector": { 4 | "state": "RUNNING", 5 | "worker_id": "my.worker.name:8083" 6 | }, 7 | "tasks": [ 8 | { 9 | "id": 0, 10 | "state": "RUNNING", 11 | "worker_id": "my.worker.name:8083" 12 | }, 13 | { 14 | "id": 2, 15 | "state": "FAILED", 16 | "worker_id": "my.worker.name:8083" 17 | }, 18 | { 19 | "id": 3, 20 | "state": "RUNNING", 21 | "worker_id": "my.worker.name:8083" 22 | }, 23 | { 24 | "id": 4, 25 | "state": "RUNNING", 26 | "worker_id": "my.worker.name:8083" 27 | } 28 | ], 29 | "type": "source" 30 | } -------------------------------------------------------------------------------- /tests/data/mocks/16-healthy-threshold-connector-s3-source.json: -------------------------------------------------------------------------------- 1 | { 2 | "name": "s3-source", 3 | "connector": { 4 | "state": "FAILED", 5 | "worker_id": "my.worker.name:8083" 6 | }, 7 | "tasks": [ 8 | { 9 | "id": 0, 10 | "state": "RUNNING", 11 | "worker_id": "my.worker.name:8083" 12 | }, 13 | { 14 | "id": 2, 15 | "state": "RUNNING", 16 | "worker_id": "my.worker.name:8083" 17 | }, 18 | { 19 | "id": 3, 20 | "state": "FAILED", 21 | "worker_id": "my.worker.name:8083" 22 | }, 23 | { 24 | "id": 4, 25 | "state": "RUNNING", 26 | "worker_id": "my.worker.name:8083" 27 | } 28 | ], 29 | "type": "source" 30 | } -------------------------------------------------------------------------------- /tests/data/mocks/16-healthy-threshold-connectors.json: -------------------------------------------------------------------------------- 1 | [ 2 | "s3-source", 3 | "s3-sink" 4 | ] -------------------------------------------------------------------------------- /tests/data/mocks/17-healthy-container-connector-connector-local-file-source.json: -------------------------------------------------------------------------------- 1 | { 2 | "name": "local-file-source", 3 | "connector": { 4 | "state": "RUNNING", 5 | "worker_id": "kafka-connect:8083" 6 | }, 7 | "tasks": [ 8 | { 9 | "id": 0, 10 | "state": "FAILED", 11 | "worker_id": "kafka-connect:8083" 12 | } 13 | ], 14 | "type": "source" 15 | } 16 | -------------------------------------------------------------------------------- /tests/data/mocks/17-healthy-container-connector-connectors.json: -------------------------------------------------------------------------------- 1 | [ 2 | "local-file-source" 3 | ] -------------------------------------------------------------------------------- /tests/data/mocks/18-healthy-container-task-connector-local-file-source.json: -------------------------------------------------------------------------------- 1 | { 2 | "name": "local-file-source", 3 | "connector": { 4 | "state": "FAILED", 5 | "worker_id": "kafka-connect:8083" 6 | }, 7 | "tasks": [ 8 | { 9 | "id": 0, 10 | "state": "RUNNING", 11 | "worker_id": "kafka-connect:8083" 12 | } 13 | ], 14 | "type": "source" 15 | } 16 | -------------------------------------------------------------------------------- /tests/data/mocks/18-healthy-container-task-connectors.json: -------------------------------------------------------------------------------- 1 | [ 2 | "local-file-source" 3 | ] -------------------------------------------------------------------------------- /tests/data/mocks/19-unhealthy-container-task-connector-local-file-source.json: -------------------------------------------------------------------------------- 1 | { 2 | "name": "local-file-source", 3 | "connector": { 4 | "state": "FAILED", 5 | "worker_id": "kafka-connect:8083" 6 | }, 7 | "tasks": [ 8 | { 9 | "id": 0, 10 | "state": "FAILED", 11 | "worker_id": "kafka-connect:8083" 12 | } 13 | ], 14 | "type": "source" 15 | } 16 | -------------------------------------------------------------------------------- /tests/data/mocks/19-unhealthy-container-task-connectors.json: -------------------------------------------------------------------------------- 1 | [ 2 | "local-file-source" 3 | ] -------------------------------------------------------------------------------- /tests/data/mocks/2-unhealthy-connector-local-file-source.json: -------------------------------------------------------------------------------- 1 | { 2 | "name": "local-file-source", 3 | "connector": { 4 | "state": "FAILED", 5 | "worker_id": "kafka-connect:8083" 6 | }, 7 | "tasks": [ 8 | { 9 | "id": 0, 10 | "state": "FAILED", 11 | "worker_id": "kafka-connect:8083" 12 | } 13 | ], 14 | "type": "source" 15 | } -------------------------------------------------------------------------------- /tests/data/mocks/2-unhealthy-connectors.json: -------------------------------------------------------------------------------- 1 | [ 2 | "local-file-source" 3 | ] -------------------------------------------------------------------------------- /tests/data/mocks/4-healthy-worker-id-correct-connector-jdbc-sink.json: -------------------------------------------------------------------------------- 1 | { 2 | "name": "jdbc-sink", 3 | "connector": { 4 | "state": "RUNNING", 5 | "worker_id": "kafka-connect:8083" 6 | }, 7 | "tasks": [ 8 | { 9 | "id": 0, 10 | "state": "RUNNING", 11 | "worker_id": "kafka-connect:8083" 12 | } 13 | ], 14 | "type": "source" 15 | } -------------------------------------------------------------------------------- /tests/data/mocks/4-healthy-worker-id-correct-connectors.json: -------------------------------------------------------------------------------- 1 | [ 2 | "jdbc-sink" 3 | ] -------------------------------------------------------------------------------- /tests/data/mocks/5-healthy-worker-id-unused-connector-jdbc-sink.json: -------------------------------------------------------------------------------- 1 | { 2 | "name": "jdbc-sink", 3 | "connector": { 4 | "state": "RUNNING", 5 | "worker_id": "some-other-worker" 6 | }, 7 | "tasks": [ 8 | { 9 | "id": 0, 10 | "state": "RUNNING", 11 | "worker_id": "some-other-worker" 12 | } 13 | ], 14 | "type": "source" 15 | } -------------------------------------------------------------------------------- /tests/data/mocks/5-healthy-worker-id-unused-connectors.json: -------------------------------------------------------------------------------- 1 | [ 2 | "jdbc-sink" 3 | ] -------------------------------------------------------------------------------- /tests/data/mocks/6-healthy-worker-id-with-other-workers-failing-connector-jdbc-sink.json: -------------------------------------------------------------------------------- 1 | { 2 | "name": "jdbc-sink", 3 | "connector": { 4 | "state": "FAILED", 5 | "worker_id": "failing-worker" 6 | }, 7 | "tasks": [ 8 | { 9 | "id": 0, 10 | "state": "FAILED", 11 | "worker_id": "failing-worker" 12 | } 13 | ], 14 | "type": "source" 15 | } -------------------------------------------------------------------------------- /tests/data/mocks/6-healthy-worker-id-with-other-workers-failing-connectors.json: -------------------------------------------------------------------------------- 1 | [ 2 | "jdbc-sink" 3 | ] -------------------------------------------------------------------------------- /tests/data/mocks/7-unhealthy-worker-id-with-other-workers-healthy-connector-jdbc-sink.json: -------------------------------------------------------------------------------- 1 | { 2 | "name": "jdbc-sink", 3 | "connector": { 4 | "state": "RUNNING", 5 | "worker_id": "test-worker" 6 | }, 7 | "tasks": [ 8 | { 9 | "id": 0, 10 | "state": "FAILED", 11 | "worker_id": "kafka-connect:8083" 12 | } 13 | ], 14 | "type": "source" 15 | } -------------------------------------------------------------------------------- /tests/data/mocks/7-unhealthy-worker-id-with-other-workers-healthy-connectors.json: -------------------------------------------------------------------------------- 1 | [ 2 | "jdbc-sink" 3 | ] -------------------------------------------------------------------------------- /tests/data/mocks/8-healthy-multiple-tasks-connector-jdbc-source.json: -------------------------------------------------------------------------------- 1 | { 2 | "name": "jdbc-source", 3 | "connector": { 4 | "state": "RUNNING", 5 | "worker_id": "my.worker.name:8083" 6 | }, 7 | "tasks": [ 8 | { 9 | "id": 0, 10 | "state": "RUNNING", 11 | "worker_id": "my.worker.name:8083" 12 | }, 13 | { 14 | "id": 2, 15 | "state": "RUNNING", 16 | "worker_id": "my.worker.name:8083" 17 | }, 18 | { 19 | "id": 3, 20 | "state": "RUNNING", 21 | "worker_id": "my.worker.name:8083" 22 | }, 23 | { 24 | "id": 4, 25 | "state": "RUNNING", 26 | "worker_id": "my.worker.name:8083" 27 | } 28 | ], 29 | "type": "source" 30 | } -------------------------------------------------------------------------------- /tests/data/mocks/8-healthy-multiple-tasks-connectors.json: -------------------------------------------------------------------------------- 1 | [ 2 | "jdbc-source" 3 | ] -------------------------------------------------------------------------------- /tests/data/mocks/9-healthy-multiple-connectors-connector-s3-sink.json: -------------------------------------------------------------------------------- 1 | { 2 | "name": "s3-sink", 3 | "connector": { 4 | "state": "RUNNING", 5 | "worker_id": "my.worker.name:8083" 6 | }, 7 | "tasks": [ 8 | { 9 | "id": 0, 10 | "state": "RUNNING", 11 | "worker_id": "my.worker.name:8083" 12 | }, 13 | { 14 | "id": 2, 15 | "state": "RUNNING", 16 | "worker_id": "my.worker.name:8083" 17 | }, 18 | { 19 | "id": 3, 20 | "state": "RUNNING", 21 | "worker_id": "my.worker.name:8083" 22 | }, 23 | { 24 | "id": 4, 25 | "state": "RUNNING", 26 | "worker_id": "my.worker.name:8083" 27 | } 28 | ], 29 | "type": "source" 30 | } -------------------------------------------------------------------------------- /tests/data/mocks/9-healthy-multiple-connectors-connector-s3-source.json: -------------------------------------------------------------------------------- 1 | { 2 | "name": "s3-source", 3 | "connector": { 4 | "state": "RUNNING", 5 | "worker_id": "my.worker.name:8083" 6 | }, 7 | "tasks": [ 8 | { 9 | "id": 0, 10 | "state": "RUNNING", 11 | "worker_id": "my.worker.name:8083" 12 | }, 13 | { 14 | "id": 2, 15 | "state": "RUNNING", 16 | "worker_id": "my.worker.name:8083" 17 | }, 18 | { 19 | "id": 3, 20 | "state": "RUNNING", 21 | "worker_id": "my.worker.name:8083" 22 | }, 23 | { 24 | "id": 4, 25 | "state": "RUNNING", 26 | "worker_id": "my.worker.name:8083" 27 | } 28 | ], 29 | "type": "source" 30 | } -------------------------------------------------------------------------------- /tests/data/mocks/9-healthy-multiple-connectors-connectors.json: -------------------------------------------------------------------------------- 1 | [ 2 | "s3-source", 3 | "s3-sink" 4 | ] -------------------------------------------------------------------------------- /tests/data/mocks/healthy-connector-details.json: -------------------------------------------------------------------------------- 1 | { 2 | } 3 | -------------------------------------------------------------------------------- /tests/mocks.py: -------------------------------------------------------------------------------- 1 | import os 2 | import socket 3 | from functools import partial 4 | from http.server import BaseHTTPRequestHandler, HTTPServer 5 | from threading import Thread 6 | 7 | 8 | # Mocks Kafka Connect REST API responses for integration tests 9 | class MockServerRequestHandler(BaseHTTPRequestHandler): 10 | 11 | def __init__(self, mock_name, *args, **kwargs): 12 | self.mock_name = mock_name 13 | super().__init__(*args, **kwargs) 14 | 15 | def do_GET(self): 16 | status_code = 503 if "unhealthy" in self.mock_name else 200 17 | try: 18 | if "auth" in self.mock_name: 19 | if "Basic" not in self.headers.get("Authorization"): 20 | self.response(401, payload="{}") 21 | 22 | if self.path == "/connectors": 23 | with open(os.path.join(os.getcwd(), "tests/data/mocks/{}-connectors.json".format(self.mock_name)), "r") as f: 24 | self.response(200, payload=f.read()) 25 | else: 26 | split_path = self.path.split("/") 27 | if len(split_path) >= 4 and split_path[3] == "status": 28 | connector_name = split_path[2] 29 | path = os.path.join(os.getcwd(), "./tests/data/mocks/{}-connector-{}.json".format(self.mock_name, connector_name)) 30 | with open(path, "r") as f: 31 | self.response(status_code, payload=f.read()) 32 | else: 33 | connector_name = split_path[2] 34 | details_status_code = 503 if "unhealthy-broker" in self.mock_name else 200 35 | path = os.path.join(os.getcwd(), 36 | "./tests/data/mocks/healthy-connector-details.json".format(self.mock_name, connector_name)) 37 | with open(path, "r") as f: 38 | self.response(details_status_code, payload=f.read()) 39 | 40 | except Exception as ex: 41 | print("Error while handling mock GET request: {}".format(ex)) 42 | self.response(500, "") 43 | 44 | def response(self, status_code, payload): 45 | self.send_response(status_code) 46 | self.send_header('Content-Type', 'application/json; charset=utf-8') 47 | self.end_headers() 48 | self.wfile.write(payload.encode('utf-8')) 49 | 50 | def log_message(self, format, *args): 51 | return 52 | 53 | 54 | def get_free_port(): 55 | s = socket.socket(socket.AF_INET, type=socket.SOCK_STREAM) 56 | s.bind(('localhost', 0)) 57 | address, port = s.getsockname() 58 | s.close() 59 | return port 60 | 61 | 62 | def start_mock_server(port, mock_name): 63 | handler = partial(MockServerRequestHandler, mock_name) 64 | mock_server = HTTPServer(('localhost', port), handler) 65 | mock_server_thread = Thread(target=mock_server.serve_forever) 66 | mock_server_thread.setDaemon(True) 67 | mock_server_thread.start() 68 | print("\nMock Kafka Connect server running on port: {}".format(port)) 69 | -------------------------------------------------------------------------------- /tests/test_server.py: -------------------------------------------------------------------------------- 1 | # tests/test_server.py 2 | 3 | import json 4 | 5 | import pytest 6 | import requests 7 | 8 | scenario_0 = ({}, "0-healthcheck-server-healthy") 9 | scenario_1 = ({"HEALTHCHECK_UNHEALTHY_STATES": "FAILED"}, "1-healthy") 10 | scenario_2 = ({"HEALTHCHECK_UNHEALTHY_STATES": "FAILED"}, "2-unhealthy") 11 | scenario_3 = ({"HEALTHCHECK_CONNECT_URL": "http://unknown-hostname:8083"}, "3-unhealthy-error") 12 | scenario_4 = ({"HEALTHCHECK_CONNECT_WORKER_ID": "kafka-connect:8083"}, "4-healthy-worker-id-correct") 13 | scenario_5 = ({"HEALTHCHECK_CONNECT_WORKER_ID": "kafka-connect:8083"}, "5-healthy-worker-id-unused") 14 | scenario_6 = ({"HEALTHCHECK_CONNECT_WORKER_ID": "kafka-connect:8083"}, "6-healthy-worker-id-with-other-workers-failing") 15 | scenario_7 = ({"HEALTHCHECK_CONNECT_WORKER_ID": "kafka-connect:8083"}, "7-unhealthy-worker-id-with-other-workers-healthy") 16 | scenario_8 = ({"HEALTHCHECK_CONNECT_WORKER_ID": "my.worker.name:8083"}, "8-healthy-multiple-tasks") 17 | scenario_9 = ({"HEALTHCHECK_CONNECT_WORKER_ID": "my.worker.name:8083"}, "9-healthy-multiple-connectors") 18 | scenario_10 = ({"HEALTHCHECK_CONNECT_WORKER_ID": "my.worker.name:8083"}, "10-unhealthy-multiple-connectors") 19 | scenario_11 = ({"HEALTHCHECK_CONNECT_WORKER_ID": "my.worker.name:8083"}, "11-healthy-no-connectors") 20 | scenario_12 = ({"HEALTHCHECK_CONNECT_WORKER_ID": "my.worker.name:8083"}, "12-unhealthy-task-with-trace") 21 | scenario_13 = ({"HEALTHCHECK_CONNECT_WORKER_ID": "unhealthy.worker:8083"}, "13-unhealthy-broker-connection") 22 | scenario_14 = ({"HEALTHCHECK_BASIC_AUTH": "username:password"}, "14-basic-auth") 23 | scenario_15 = ({"HEALTHCHECK_FAILURE_THRESHOLD_PERCENTAGE": "10"}, "15-unhealthy-threshold") 24 | scenario_16 = ({"HEALTHCHECK_FAILURE_THRESHOLD_PERCENTAGE": "50"}, "16-healthy-threshold") 25 | scenario_17 = ({"HEALTHCHECK_CONSIDERED_CONTAINERS": "CONNECTOR"}, "17-healthy-container-connector") 26 | scenario_18 = ({"HEALTHCHECK_CONSIDERED_CONTAINERS": "TASK"}, "18-healthy-container-task") 27 | scenario_19 = ({"HEALTHCHECK_CONSIDERED_CONTAINERS": "TASK"}, "19-unhealthy-container-task") 28 | other_scenarios = ({}, None) 29 | 30 | 31 | @pytest.mark.parametrize("run_backend", [scenario_0], indirect=True) 32 | def test_0_healthcheck_server_healthy(run_backend): 33 | with open("tests/data/expected/0-healthcheck-server-healthy.json", "r") as f: 34 | response = requests.get("http://localhost:18083/ping") 35 | assert response.status_code == 200 36 | assert json.loads(response.content.decode("utf-8")) == json.load(f) 37 | 38 | 39 | @pytest.mark.parametrize("run_backend", [scenario_1], indirect=True) 40 | def test_1_healthy(run_backend, request): 41 | with open("tests/data/expected/1-healthy.json", "r") as f: 42 | response = requests.get("http://localhost:18083") 43 | logs = request.config.cache.get("logs", "") == "" 44 | assert response.status_code == 200 45 | assert json.loads(response.content.decode("utf-8")) == json.load(f) 46 | 47 | 48 | @pytest.mark.parametrize("run_backend", [scenario_2], indirect=True) 49 | def test_2_unhealthy(run_backend): 50 | with open("tests/data/expected/2-unhealthy.json", "r") as f: 51 | response = requests.get("http://localhost:18083") 52 | assert response.status_code == 503 53 | assert json.loads(response.content.decode("utf-8")) == json.load(f) 54 | 55 | 56 | @pytest.mark.parametrize("run_backend", [scenario_3], indirect=True) 57 | def test_3_unhealthy_error(run_backend): 58 | with open("tests/data/expected/3-unhealthy-error.json", "r") as f: 59 | response = requests.get("http://localhost:18083/") 60 | expected_response = json.load(f) 61 | actual_response = json.loads(response.content.decode("utf-8")) 62 | assert response.status_code == 503 63 | assert actual_response["healthy"] == expected_response["healthy"] 64 | assert actual_response["message"] == expected_response["message"] 65 | assert actual_response["failure_states"] == expected_response["failure_states"] 66 | assert expected_response["error"] in actual_response["error"] 67 | 68 | 69 | @pytest.mark.parametrize("run_backend", [scenario_4], indirect=True) 70 | def test_4_healthy_worker_id_correct(run_backend): 71 | with open("tests/data/expected/4-healthy-worker-id-correct.json", "r") as f: 72 | response = requests.get("http://localhost:18083") 73 | assert response.status_code == 200 74 | assert json.loads(response.content.decode("utf-8")) == json.load(f) 75 | 76 | 77 | @pytest.mark.parametrize("run_backend", [scenario_5], indirect=True) 78 | def test_5_healthy_worker_id_unused(run_backend): 79 | with open("tests/data/expected/5-healthy-worker-id-unused.json", "r") as f: 80 | response = requests.get("http://localhost:18083") 81 | assert response.status_code == 200 82 | assert json.loads(response.content.decode("utf-8")) == json.load(f) 83 | 84 | 85 | @pytest.mark.parametrize("run_backend", [scenario_6], indirect=True) 86 | def test_6_healthy_worker_id_with_other_workers_failing(run_backend): 87 | with open("tests/data/expected/6-healthy-worker-id-with-other-workers-failing.json", "r") as f: 88 | response = requests.get("http://localhost:18083") 89 | assert response.status_code == 200 90 | assert json.loads(response.content.decode("utf-8")) == json.load(f) 91 | 92 | 93 | @pytest.mark.parametrize("run_backend", [scenario_7], indirect=True) 94 | def test_7_unhealthy_worker_id_with_other_workers_healthy(run_backend): 95 | with open("tests/data/expected/7-unhealthy-worker-id-with-other-workers-healthy.json", "r") as f: 96 | response = requests.get("http://localhost:18083") 97 | assert response.status_code == 503 98 | assert json.loads(response.content.decode("utf-8")) == json.load(f) 99 | 100 | 101 | @pytest.mark.parametrize("run_backend", [scenario_8], indirect=True) 102 | def test_8_healthy_multiple_tasks(run_backend): 103 | with open("tests/data/expected/8-healthy-multiple-tasks.json", "r") as f: 104 | response = requests.get("http://localhost:18083") 105 | assert response.status_code == 200 106 | assert json.loads(response.content.decode("utf-8")) == json.load(f) 107 | 108 | 109 | @pytest.mark.parametrize("run_backend", [scenario_9], indirect=True) 110 | def test_9_healthy_multiple_connectors(run_backend): 111 | with open("tests/data/expected/9-healthy-multiple-connectors.json", "r") as f: 112 | response = requests.get("http://localhost:18083") 113 | assert response.status_code == 200 114 | assert json.loads(response.content.decode("utf-8")) == json.load(f) 115 | 116 | 117 | @pytest.mark.parametrize("run_backend", [scenario_10], indirect=True) 118 | def test_10_unhealthy_multiple_connectors(run_backend): 119 | with open("tests/data/expected/10-unhealthy-multiple-connectors.json", "r") as f: 120 | response = requests.get("http://localhost:18083") 121 | assert response.status_code == 503 122 | assert json.loads(response.content.decode("utf-8")) == json.load(f) 123 | 124 | 125 | @pytest.mark.parametrize("run_backend", [scenario_11], indirect=True) 126 | def test_11_healthy_no_connectors(run_backend): 127 | with open("tests/data/expected/11-healthy-no-connectors.json", "r") as f: 128 | response = requests.get("http://localhost:18083") 129 | assert response.status_code == 200 130 | assert json.loads(response.content.decode("utf-8")) == json.load(f) 131 | 132 | 133 | @pytest.mark.parametrize("run_backend", [scenario_12], indirect=True) 134 | def test_12_unhealthy_task_with_trace(run_backend): 135 | with open("tests/data/expected/12-unhealthy-task-with-trace.json", "r") as f: 136 | response = requests.get("http://localhost:18083") 137 | assert response.status_code == 503 138 | assert json.loads(response.content.decode("utf-8")) == json.load(f) 139 | 140 | 141 | @pytest.mark.parametrize("run_backend", [scenario_13], indirect=True) 142 | def test_13_unhealthy_broker_connection(run_backend): 143 | with open("tests/data/expected/13-unhealthy-broker-connection.json", "r") as f: 144 | response = requests.get("http://localhost:18083") 145 | assert response.status_code == 503 146 | assert json.loads(response.content.decode("utf-8")) == json.load(f) 147 | 148 | 149 | @pytest.mark.parametrize("run_backend", [scenario_14], indirect=True) 150 | def test_14_basic_auth(run_backend): 151 | with open("tests/data/expected/14-basic-auth.json", "r") as f: 152 | response = requests.get("http://localhost:18083") 153 | assert response.status_code == 200 154 | assert json.loads(response.content.decode("utf-8")) == json.load(f) 155 | 156 | 157 | @pytest.mark.parametrize("run_backend", [scenario_15], indirect=True) 158 | def test_15_unhealthy_threshold(run_backend): 159 | with open("tests/data/expected/15-unhealthy-threshold.json", "r") as f: 160 | response = requests.get("http://localhost:18083") 161 | assert response.status_code == 503 162 | assert json.loads(response.content.decode("utf-8")) == json.load(f) 163 | 164 | 165 | @pytest.mark.parametrize("run_backend", [scenario_16], indirect=True) 166 | def test_16_healthy_threshold(run_backend): 167 | with open("tests/data/expected/16-healthy-threshold.json", "r") as f: 168 | response = requests.get("http://localhost:18083") 169 | assert response.status_code == 200 170 | assert json.loads(response.content.decode("utf-8")) == json.load(f) 171 | 172 | 173 | @pytest.mark.parametrize("run_backend", [scenario_17], indirect=True) 174 | def test_17_healthy_container_connector(run_backend): 175 | with open("tests/data/expected/17-healthy-container-connector.json", "r") as f: 176 | response = requests.get("http://localhost:18083") 177 | assert response.status_code == 200 178 | assert json.loads(response.content.decode("utf-8")) == json.load(f) 179 | 180 | 181 | @pytest.mark.parametrize("run_backend", [scenario_18], indirect=True) 182 | def test_18_healthy_container_task(run_backend): 183 | with open("tests/data/expected/18-healthy-container-task.json", "r") as f: 184 | response = requests.get("http://localhost:18083") 185 | assert response.status_code == 200 186 | assert json.loads(response.content.decode("utf-8")) == json.load(f) 187 | 188 | 189 | @pytest.mark.parametrize("run_backend", [scenario_19], indirect=True) 190 | def test_19_unhealthy_container_task(run_backend): 191 | with open("tests/data/expected/19-unhealthy-container-task.json", "r") as f: 192 | response = requests.get("http://localhost:18083") 193 | assert response.status_code == 503 194 | assert json.loads(response.content.decode("utf-8")) == json.load(f) 195 | 196 | 197 | @pytest.mark.parametrize("test_input", ["test", "does/not/exist", "kafka", "this-is-a-long/url", "connectors"]) 198 | @pytest.mark.parametrize("run_backend", [other_scenarios], indirect=True) 199 | def test_for_404s(run_backend, test_input): 200 | response = requests.get("http://localhost:18083/{}".format(test_input)) 201 | assert response.status_code == 404 202 | assert response.content == b"" 203 | 204 | 205 | @pytest.mark.parametrize("test_input", ["", "test", "does/not/exist"]) 206 | @pytest.mark.parametrize("run_backend", [other_scenarios], indirect=True) 207 | def test_head_request(run_backend, test_input): 208 | response = requests.head("http://localhost:18083/{}".format(test_input)) 209 | assert response.status_code == 200 210 | assert response.headers.get("Content-type") == "text/html" 211 | assert response.content == b"" 212 | -------------------------------------------------------------------------------- /tests/test_version.py: -------------------------------------------------------------------------------- 1 | import pytest 2 | 3 | from .context import kafka_connect_healthcheck 4 | 5 | test_data = [ 6 | "0.0.1", 7 | ] 8 | 9 | 10 | @pytest.mark.parametrize("test_input", test_data) 11 | def test_get_version(test_input): 12 | old_version = kafka_connect_healthcheck.version.__version__ 13 | kafka_connect_healthcheck.version.__version__ = test_input 14 | assert kafka_connect_healthcheck.version.get_version() == test_input 15 | kafka_connect_healthcheck.version.__version__ = old_version 16 | -------------------------------------------------------------------------------- /tox.ini: -------------------------------------------------------------------------------- 1 | [tox] 2 | envlist = py35,py36,py37,py38 3 | 4 | [gh-actions] 5 | python = 6 | 3.5: py35 7 | 3.6: py36 8 | 3.7: py37 9 | 3.8: py38 10 | 11 | [testenv] 12 | deps = 13 | pytest 14 | pytest-cov 15 | mock 16 | flake8 17 | flake8-quotes 18 | requests 19 | coverage 20 | commands = 21 | flake8 --max-line-length=160 --inline-quotes '"' --exclude=.tox,.git,htmlcov,build,dist,tests,docs,venv . 22 | pytest 23 | --------------------------------------------------------------------------------