Dynamic hot spares based on running pipelines

[sgleske-ias]

What feature do you want to see added?

In general, I may know exactly how many hot spares are required based on currently running pipelines. It would be nice to have a dynamic hot spare logic which is based on actual running pipelines instead of a static value.

Feature behavior:

• per running pipeline hot spares gets 1 * <running pipelines> * <user configurable multiplier>
• Configurable hot spares maximum to prevent spares getting larger than a specified value.
• Configurable hot spare idle timeout where hot spares is re-evaluated. e.g. if no running pipelines in 5 minutes I want hot spares to be 0.

Example: I would like 2 hot spares per running Jenkins pipeline. However, I do not want hot spares to go above 10 hot spares as its absolute max.

Benefits:

• My Jenkins infrastructure would be able to scale down to 0 ec2 instances when no work needs to be done.
• Reasonable hot spare capacity can be determined by Jenkins administrator based on self-designed architecture or data from APM monitoring Jenkins.

Upstream changes

Unsure

Are you interested in contributing this feature?

I am willing to fully implement this feature. My open source handle is @samrocketman where I maintain some plugins and have contributed patches across the Jenkins project since 2011.