Uploaded image for project: 'Couchbase Server'
  1. Couchbase Server
  2. MB-56517

Report raw stats instead of computed stats

    XMLWordPrintable

Details

    • 0

    Description

      As seen in MB-56483 ns_server computes "rate" stats using information from sigar as well as elapsed times in ns_server. Because of this the rates may end up "fuzzy" (> 100%). As prometheus provides a rich set of formulas the stats should be reported to prometheus as raw stats instead of computed stats.

       

       

      Issue Resolution
      The Cluster Manager’s computed utilization stats were inaccurate due to time interval discrepancies in components where data was collected. The Cluster Manager now reports raw stats as Prometheus counters.

      Attachments

        1. AWSLinuxOldNewStat.png
          73 kB
          Steve Watanabe
        2. screenshot-1.png
          206 kB
          Steve Watanabe
        3. screenshot-2.png
          71 kB
          Steve Watanabe
        4. Screenshot 2023-07-18 at 12.23.32 PM.png
          323 kB
          Pulkit Matta
        5. Screenshot 2023-07-18 at 12.23.51 PM.png
          275 kB
          Pulkit Matta
        6. Screenshot 2023-07-18 at 12.24.14 PM.png
          464 kB
          Pulkit Matta
        7. WindowsStats.png
          463 kB
          Steve Watanabe

        Issue Links

          For Gerrit Dashboard: MB-56517
          # Subject Branch Project Status CR V

          Activity

            People

              pulkit.matta Pulkit Matta
              steve.watanabe Steve Watanabe
              Votes:
              0 Vote for this issue
              Watchers:
              6 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved:

                PagerDuty