Scylla Monitoring Stack 3.2

The Scylla team is pleased to announce the release of Scylla Monitoring Stack 3.2.

Scylla Monitoring Stack is an open-source stack for monitoring Scylla Enterprise and Scylla Open Source, based on Prometheus and Grafana. Scylla Monitoring Stack 3.2 supports:

  • Scylla Open Source versions 3.1, 3.2 and the upcoming 3.3
  • Scylla Enterprise versions 2018.x and 2019.x
  • Scylla Manager 1.4.x and Scylla Manager 2.0.x

Related Links

New in Scylla Monitoring Stack 3.2

  • Add annotation for Scylla Manager tasks – The manager dashboard has now annotations for the beginning and termination of its tasks (repair and backup)

  • Show all possible I/O queue classes, the IO dashboard is now split into sections, each containing all possible io queue classes. This will support dynamic classes that are being added to the Enterprise release.

  • New Scylla Open Source 3.3 dashboards
  • Move to Grafana 6.6. More on Grafana release can be found here
  • Move to Prometheus 2.15.1. More on Prometheus release can be found here
  • Grafana security – Support specifying the default role. This will allow you to start the monitoring stack with a restricted default role.
    For example: start-all.sh -Q Viewer will start Grafana with the default role set to Viewer.
    This is useful when your dashboards are open to a public audience, in which you prefer that the users will not be able to make changes to the dashboard.
  • A new BYPASS CACHE panel in the CQL dashboard, showing the number of queries that uses BYPASS CACHE

  • New Hinted Handoff panel in the detailed dashboard.
  • A new annotation for reactor stalls. Scylla reactors report any stalles larger than a pre set value (command line parameter blocked-reactor-notify-ms, default 100ms)

  • A new ignore future metrics to the Error dashboard.

Bug Fixes

  • Bring back help strings that removed by mistake #841
  • Compaction share graph shows wrong results #817

19 Mar 2020