Uploaded image for project: 'Red Hat 3scale API Management'
  1. Red Hat 3scale API Management
  2. THREESCALE-8717

Provide granular control of PDBs when not using specific pods

    XMLWordPrintable

Details

    • Feature Request
    • Resolution: Can't Do
    • Major
    • None
    • 2.12.0 GA
    • 3scale Operator
    • False
    • None
    • False
    • Not Started
    • Not Started
    • Not Started
    • Not Started
    • Not Started
    • Not Started
    • 0
    • 0% 0%
    • RHOAM Sprint 47

    Description

      When setting PDB in APIManager, as described in our documentation it applies for all 3scale components. 

      $ oc get pdb
      NAME                 MIN AVAILABLE   MAX UNAVAILABLE   ALLOWED DISRUPTIONS   AGE
      apicast-production   N/A             1                 1                     4m14s
      apicast-staging      N/A             1                 1                     4m14s
      backend-cron         N/A             1                 1                     4m15s
      backend-listener     N/A             1                 1                     4m15s
      backend-worker       N/A             1                 1                     4m15s
      system-app           N/A             1                 1                     4m15s
      system-sidekiq       N/A             1                 1                     4m15s
      zync                 N/A             1                 1                     4m14s
      zync-que             N/A             1                 1                     4m14s
      

      If some components are not used, for example zync and zync-que the allowed disruptions are set to 0:

      $ oc get pdb
      NAME                 MIN AVAILABLE   MAX UNAVAILABLE   ALLOWED DISRUPTIONS   AGE
      apicast-production   N/A             1                 1                     5m42s
      apicast-staging      N/A             1                 1                     5m42s
      backend-cron         N/A             1                 1                     5m43s
      backend-listener     N/A             1                 1                     5m43s
      backend-worker       N/A             1                 1                     5m43s
      system-app           N/A             1                 1                     5m43s
      system-sidekiq       N/A             1                 1                     5m43s
      zync                 N/A             1                 0                     5m42s
      zync-que             N/A             1                 0                     5m42s
      
      

      When a customer defines an alert based on allowed disruptions = 0 to prevent service outages and the service is scaled to 0 in purpose it creates a false positive alarm. The false alarm blocks updates on the cluster.

      The request is to add more granular control of PDB in 3scale, per component level instead of application level.

      Attachments

        Issue Links

          Activity

            People

              Unassigned Unassigned
              rhn-support-avilatus Anna Vila Tusell
              Votes:
              1 Vote for this issue
              Watchers:
              5 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: