Uploaded image for project: 'Couchbase Monitoring and Observability Stack'
  1. Couchbase Monitoring and Observability Stack
  2. CMOS-101

Prometheus discovery: trying multiple ports is expensive

    XMLWordPrintable

Details

    • Improvement
    • Resolution: Unresolved
    • Major
    • 1.0
    • None
    • cluster-monitor
    • None

    Description

      In order for prometheus->cbmultimanager discovery to work for both 6.x and 7.x clusters, we need to try to connect to each node at most three times - once with the port that Prometheus gives us (which may be 8091 on 7.x, 9091 on 6.x with the Exporter, or something else if custom ports are in use), and once each with 8091 and 18091 (and possibly more once we support custom ports). This is a fairly expensive operation, made more expensive by the fact that we need to try it for each node we discover, as they might all be from different clusters but we have no way of knowing that short of getting the cluster UUID.

      Investigate more efficient ways of doing this.

      Attachments

        Issue Links

          No reviews matched the request. Check your Options in the drop-down menu of this sections header.

          Activity

            People

              Unassigned Unassigned
              marks.polakovs Marks Polakovs (Inactive)
              Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

              Dates

                Created:
                Updated:

                Gerrit Reviews

                  There are no open Gerrit changes

                  PagerDuty