You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Case 2 will be especially painful, when you run pipelines that deploy specifically named branches for product versions to critical target environments, like qa or production systems.
You simply want to know when a critical system is running a deployment. Hence, getting the right notifications to target audience.
example:
Dedicated multibranch pipeline job is deploying artifacts to remote systems. source is a "named" release branch of that job.
In this case it is very likely that the release branch has never been seen by the pipeline. Due to this bug, no one gets notified of that process starting.
The text was updated successfully, but these errors were encountered:
The very first "started" notification is only received, when manually creating a pipeline job and additionally configuring the WebHook manually in the job configuration UI.
The very first "started" notification is not received, when creating a pipelione job and let the pipeline script define the WebHook via the "options" section.
The Bug occurs for Pipelines via SCM (where the Jenkinsfile configures the job options, triggers) and in Multibranch-Pipeline Jobs for the same reason. After parsing the Jenkinsfile and adding the webhook for the very first time the "started" notification is neglected. If it was configured before (by a previous run or manually) it will be send out.
You may now argue the Job was already started,
triggered Build
read config
read Jenkinsfile
parsed Jenkinsfile
add new web-hook options as defined in Jenkinsfile
run rest of the pipeline
which may technically correct.
But this will still be an issue where the web-hook Options are not manually configured and rather read from scm/Jenkinsfile as in Multibranch Pipelines.
OK, so based on my experiences with Jenkinsfile and the way how this work when loading it from git repository this is not the plugin issue but the configuration of your job. I believe that for the first run you do not have configured hooks and that's why they are not executed.
Then, when job is running - not only first time but also third and fourth and so on - job configuration is updated based on what you have in your Jenkinsfile. This is by Jenkins design.
You can probably prove it by comparing config.xml version after the project is setup and after the first execution is completed. Look for options/properties section to see what has changed after the job execution.
Or similar: You have your hook configured and then you decided that they should be removed. Do this in git repository and check what happens for the next job run: will the notification be sent because you have it OR will not be sent because you have decided to remove it
Jenkins and plugins versions report
Environment
What Operating System are you using (both controller, and any agents involved in the problem)?
agent: ubuntu:20.04
controller: docker jenkins/jenkins:lts-jdk11
Reproduction steps
case 1, simple pipeline:
case 2: multibranch pipeline:
Expected Results
case 1, simple pipeline:
case 2: multibranch pipeline:
Actual Results
case 1, simple pipeline:
case 2: multibranch pipeline:
Anything else?
Case 2 will be especially painful, when you run pipelines that deploy specifically named branches for product versions to critical target environments, like qa or production systems.
You simply want to know when a critical system is running a deployment. Hence, getting the right notifications to target audience.
example:
Dedicated multibranch pipeline job is deploying artifacts to remote systems. source is a "named" release branch of that job.
In this case it is very likely that the release branch has never been seen by the pipeline. Due to this bug, no one gets notified of that process starting.
The text was updated successfully, but these errors were encountered: