Restart of system that are killed due to OOM

Completed

Comments

2 comments

  • Official comment
    Geir Ove Grønmo

    We are considering adding support enforcing an upper-limit on the memory and cpu allocations of microservices based on subscription size.

    Comment actions Permalink
  • Geir Ove Grønmo

    In the new Kubernetes clusters the microservices are being automatically restarted when they are OOM-killed. There is a back-off scheme active which will mean that it takes subsequently longer time between restarts if it happens several times.

    0
    Comment actions Permalink

Please sign in to leave a comment.