Hello,

we are currently facing issues where our agent pods (created when build order triggers its creation) get killed with reason being OOM. 

This can be fixed by setting memory limit to 3072.

However, this setting only persists to next restart of Jenkins, then its back to its default value of 1024.

Furthermore, memory limit value is not even taken from helm values.yaml, which we are using to install the jenkins (it gets ignored)

Please advise