I have a Jenkins instance that uses the mesos plugin . Almost all of my tasks run through Mesos tasks. I would like to make the working generation a little more aggressive.
The current problem is that for the mesos plugin, I have all the tasks that designate mesos tasks as one-time service slaves, and when assembly is performed on one of these slaves, Jenkins makes any task in the queue wait for a potential executor on these subordinates, unlike rotation of new instances.
Based on the logs, it also seems that Jenkins has a timer that periodically checks whether any subordinates should be deployed based on # jobs in the queue or excessive workload. Is it possible to reduce the polling interval for this process?
K-ran source share