Kafka Connect Connector for Jenkins Open Source Continuous Integration Tool.
kafka-connect-jenkins is a Kafka Connector for loading data from Jenkins Open Source Continuous Integration Tool.
+--------+ 0..* +-------+ 0..* +-------+ 1 +--------------+
| Server +--------> | Job +-------> | Build +--------> | BuildDetails |
+--------+ +-------+ +-------+ +--------------+
The following comparison might give you more insight.
Example | Input Partitions (Logical) | Records | Offsets/Position |
---|---|---|---|
Set of files | each File | each Line | line Number within a file |
Database instance | each Table | each Row | primary id + timestamp |
Jenkins Server | each Job | BuildDetails of each Build (lastBuild) | build number of lastBuild |
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"
}
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.
Property | Description | Required? | Default value |
---|---|---|---|
connector.class |
Class implementing source connector for Jenkins. | Yes | org.aravind.oss.kafka.connect. jenkins.JenkinsSourceConnector |
tasks.max |
Yes | 1 | |
jenkins.base.url |
The URL where jenkins server is running | Yes | None |
jenkins.pollIntervalInMillis |
Frequency (in milliseconds) to poll for new build events in Jenkins | Yes | 1 minute |
jenkins.username |
If your Jenkins is secured, you can provide the username with this property | No | None |
jenkins.password.or.api.token |
If your Jenkins is secured, you can provide the password or api token with this property | No | None |
jenkins.connection.timeoutInMillis |
Connection timeout in milliseconds. This denotes the time elapsed before the connection established or Server responded to connection request. | Yes | 500 |
jenkins.read.timeoutInMillis |
Response read timeout in milliseconds. After establishing the connection, the client socket waits for response after sending the request. This is the elapsed time since the client has sent request to the server before server responds. | Yes | 3000 |
jenkins.jobs.resource.path |
Relative path to REST API | Yes | /api/json |
topic |
Name of the topic where the Build status records are written to. Make sure you explicitly create this topic using tools provided by Kafka. Do not rely on the default topic creation functionality in PRODUCTION. | Yes | jenkins.connector.topic |
While testing, you might want to run the connector in standalone mode. Follow these steps
- Download Confluent platform 3.0.0 ZIP file from here
- Unzip it
unzip confluent-3.0.0-2.11.zip
- Move to Confluent home directory
cd confluent-3.0.0
- Start Zookeeper
./bin/zookeeper-server-start etc/kafka/zookeeper.properties
- Start Kafka Broker
./bin/kafka-server-start ./etc/kafka/server.properties
- Clone https://github.com/yaravind/kafka-connect-jenkins.git
- Run
mvn clean install
- Update the
jenkins.base.url
property inconnect-jenkins-source.properties
file with your jenkins base url - 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=...
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.
Intent | Configuration (DEBUG or TRACE ) |
---|---|
Enable logging for Jenkins Source Connector only | log4j.logger.org.aravind.oss.kafka.connect.jenkins.JenkinsSourceConnector=DEBUG log4j.logger.org.aravind.oss.kafka.connect.lib.TaskConfigBuilder=DEBUG |
Enable logging for Jenkins Source Task only | log4j.logger.org.aravind.oss.kafka.connect.jenkins.JenkinsSourceTask=DEBUG log4j.logger.org.aravind.oss.kafka.connect.jenkins.ReadYourWritesOffsetStorageAdapter=ERROR log4j.logger.org.aravind.oss.kafka.connect.lib.Partitions=DEBUG |
Enable logging for communication with Jenkins API only | log4j.logger.org.aravind.oss.jenkins.JenkinsClient=TRACE |
Replace
localhost
with your server and8083
withrest.port
configuration value.
REST API | Details |
---|---|
http://localhost:8083/ |
Kafka Connector |
http://localhost:8083/connectors/ |
All deployed connectors |
http://localhost:8083/connectors/kafka-jenkins-source-connector/ |
JenkinsSourceConnector |
http://localhost:8083/connectors/kafka-jenkins-source-connector/config |
JenkinsSourceConnector config |
http://localhost:8083/connectors/kafka-jenkins-source-connector/tasks |
JenkinsSourceConnector tasks |
- 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.
- 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
- Source code: https://github.com/yaravind/kafka-connect-jenkins
- Issue tracker: https://github.com/yaravind/kafka-connect-jenkins/issues
The project is licensed under the Apache 2 license.