Kafka Connect Jenkins

Kafka Connect Connector for Jenkins Open Source Continuous Integration Tool.

What is it?

kafka-connect-jenkins is a Kafka Connector for loading data from Jenkins Open Source Continuous Integration Tool.

Details

+--------+   0..*   +-------+  0..*   +-------+   1      +--------------+
| Server +--------> |  Job  +-------> | Build +--------> | BuildDetails |
+--------+          +-------+         +-------+          +--------------+

The following comparison might give you more insight.

Example

We will use Apache Jenkins REST API to demonstrate an example.

Resource	Details
Get list of all Jobs: https://builds.apache.org/api/json GET lastBuild of Abdera-trunk Job: https://builds.apache.org/job/Abdera-trunk/api/json GET BuildDetails for build 2546 of job Abdera-trunk: https://builds.apache.org/job/Abdera-trunk/2546/api/json

What gets published to the topic?

The BuildDetails is treated as an event and is persisted to the topic in JSON format. Below is the content of a sample event

{
    "actions": [
        {
            "causes": [
                {
                    "shortDescription": "Started by an SCM change"
                }
            ]
        },
        {},
        {},
        {},
        {},
        {},
        {},
        {
            "failCount": 0,
            "skipCount": 1,
            "totalCount": 491,
            "urlName": "testReport"
        },
        {},
        {}
    ],
    "artifacts": [],
    "building": false,
    "description": null,
    "displayName": "#2546",
    "duration": 480582,
    "estimatedDuration": 457794,
    "executor": null,
    "fullDisplayName": "Abdera-trunk #2546",
    "id": "2015-08-25_22-08-43",
    "keepLog": false,
    "number": 2546,
    "queueId": -1,
    "result": "SUCCESS",
    "timestamp": 1440540523000,
    "url": "https://builds.apache.org/job/Abdera-trunk/2546/",
    "builtOn": "jenkins-ubuntu-1404-4gb-c51",
    "changeSet": {
        "items": [
            {
                "affectedPaths": [
                    "client/src/test/java/org/apache/abdera/test/client/util/MultipartRelatedRequestEntityTest.java"
                ],
                "author": {
                    "absoluteUrl": "https://builds.apache.org/user/veithen",
                    "fullName": "veithen"
                },
                "commitId": "1697770",
                "timestamp": 1440537458260,
                "date": "2015-08-25T21:17:38.260515Z",
                "msg": "Use factory to create FOMEntry instance.",
                "paths": [
                    {
                        "editType": "edit",
                        "file": "/abdera/java/trunk/client/src/test/java/org/apache/abdera/test/client/util/MultipartRelatedRequestEntityTest.java"
                    }
                ],
                "revision": 1697770,
                "user": "veithen"
            }
        ],
        "kind": "svn",
        "revisions": [
            {
                "module": "https://svn.apache.org/repos/asf/abdera/java/trunk",
                "revision": 1697770
            }
        ]
    },
    "culprits": [
        {
            "absoluteUrl": "https://builds.apache.org/user/veithen",
            "fullName": "veithen"
        }
    ],
    "mavenArtifacts": {},
    "mavenVersionUsed": "3.0.4"
}

What can I do with it?

In an organization, there can be one or more CI/CD servers (usually Jenkins) managing the day to day builds and deployments of various components or services. We usually see each business unit (or vertical) managing their own Jenkins instance. Any such organization can benefit from treating build events same like any other transactional or master data. Here are some use-cases that these build events can enable.

Configurations

How to use it?

Standalone mode

While testing, you might want to run the connector in standalone mode. Follow these steps

1. Download Confluent platform 3.0.0 ZIP file from here
2. Unzip it unzip confluent-3.0.0-2.11.zip
3. Move to Confluent home directory cd confluent-3.0.0
4. Start Zookeeper ./bin/zookeeper-server-start etc/kafka/zookeeper.properties
5. Start Kafka Broker ./bin/kafka-server-start ./etc/kafka/server.properties
6. Clone https://github.com/yaravind/kafka-connect-jenkins.git
7. Run mvn clean install
8. Update the jenkins.base.url property in connect-jenkins-source.properties file with your jenkins base url
9. Add Jenkins connector to classpath and start the connector in standalone mode CLASSPATH=./target/kafka-connect-jenkins-0.5.0-SNAPSHOT-package/share/java/kafka-connect-jenkins/* <<path-to-confluent-home-directory>>/bin/connect-standalone connect-standalone.properties connect-jenkins-source.properties

If you need to proxy to connect to Jenkins then append these to the above command -Dhttp.proxyHost=... -Dhttp.proxyPort=... -Dhttps.proxyHost=... -Dhttps.proxyPort=...

Logging

You can enable the logging for the connector by adding log4j.logger.org.aravind.oss=DEBUG (TRACE) to connect-log4j.properties. Following table can help you with finer level of logging.

Monitoring

Replace localhost with your server and 8083 with rest.port configuration value.

Limitations

• Saves only the most recent build (lastBuild) know after configured jenkins.pollIntervalInMillis. i.e. if a Job has been built multiple times within the poll intervals, it isn't accounted for.
• Requires JDK 8 to run the connector. Making JDK 7 compatible version isn't a big deal. Raise an issue if you need one.

Dependencies

• Maven 3.x
• JDK 1.8 or newer
• Spock Framework
• Snap CI for Continuous Integration
• JaCoCo and Coveralls for code coverage
• Codacy to measure code quality

Contribute

• Source code: https://github.com/yaravind/kafka-connect-jenkins
• Issue tracker: https://github.com/yaravind/kafka-connect-jenkins/issues

License

The project is licensed under the Apache 2 license.