How to disable Service check when host (via ssh) is down ?

This forum was archived to /woltlab and is now in read-only mode. Please register a new account on our new community platform.

You can create a thread on the new site and link to an archived thread. This archive is available as knowledge base, safe and secured.

More details here.
  • Hi ,

    I have configured 100 servers with agentless setup (ssh) it's pretty annoying to see service check notifications on web dashboard when a server is down , How can I hide service check alerts from icinga web when a host is down (no ssh access)?

  • This has been asked in several threads already, I've created a Github issue to enhance the docs (the linked URLs should contain solutions).

  • There are two things here:

    1. Notifications

    There is an implicit dependency from a service to its host that prevents that notifications are sent out for services if their host is down.

    2. Service checks

    The default with above dependency is that they continue to run.

    You can create an explicit dependency that disables these checks.

    There are two things to take into account:

    It might be helpfull to have the host check interval 2-3 times shorter than the service check intervals to reduce situations where the host fails

    but is not checked yet. If the (now failing) services are checked before the host, you still get service down notifications because above dependency does not trigger until the host is checked to be down.

    Disabling service checks does result in "freezing" the last known service state in the Web Frontend. It will *not* turn these into "unknown".

    If that would be your goal, an event commad configured to the host could run a script that changes all service states of that host to unknown via an api call.

    However, this approach turns out to be problematic when remote satellites are writing a backlog because they have lost the connection to their parent / master:

    Once the master is reachable again, this backlog can not be replayed anymore if you forcibly insert status changes via the API.

    So make your decision...