For Jenkins Essentials, we are going to very often need to determine how healthy a Jenkins instance is.
At least for the data Snapshotting/rollback (JENKINS-49406), we will need to quickly decide if we trigger a rollback or not.
The health is probably going to be a mix of several aspects, among others, for instance:
- is Jenkins answering
- are there (more) warning, or error logs than usual? Overall, did the amount of logs explode?
- can a build run?
- ...
We need to think carefully about that I think, and write the associated JEP for review/feedback.
- relates to
-
JENKINS-49406 Design (JEP) the Evergreen snapshotting data safety system
-
- Resolved
-
-
-
- Resolved
-
-
-
- Closed
-
- links to
[JENKINS-50294] Design (JEP) Essentials Instance Client Health Checking
| Link |
New:
This issue relates to |
| Description |
Original:
For Jenkins Essentials, we are going to very often need to determine how healthy a Jenkins instance is. At least for the data Snapshotting/rollback ( The _health_ is probably going to be a mix of several aspects, among others, for instance: * is Jenkins answering * are there (more) warning, or error logs than usual? Overall, did the amount of logs explode? * can a build run? * ... We need to think carefully about that I think, and write the associated JEP for review/feedback. |
New:
For Jenkins Essentials, we are going to very often need to determine how healthy a Jenkins instance is. At least for the data Snapshotting/rollback ( The _health_ is probably going to be a mix of several aspects, among others, for instance: * is Jenkins answering * are there (more) warning, or error logs than usual? Overall, did the amount of logs explode? * can a build run? * ... We need to think carefully about that I think, and write the associated JEP for review/feedback. |
| Assignee | Original: R. Tyler Croy [ rtyler ] | New: Baptiste Mathus [ batmat ] |
| Link |
New:
This issue relates to |
| Status | Original: Open [ 1 ] | New: In Progress [ 3 ] |
| Rank | New: Ranked higher |
| Remote Link | New: This issue links to "Health checking prototype PR (Web Link)" [ 20375 ] |
So I have tried a bit configuring the equivalent of the metrics XML file using CasC to no avail. So I filed https://github.com/jenkinsci/configuration-as-code-plugin/issues/169 to get feedback from the plugin developers to save some cycles (either it's possible, and they'll tell me what I missed, or it's not and I guess we'll work together to have the metrics plugin supported by casc).
https://github.com/jenkins-infra/evergreen/pull/44
Demo that checking the HTTP status code of the /login URL, and configuring the metrics plugin to expose a jenkinsURL/metrics/evergreen/health-check URL are doable.