Scylla Monitoring Stack 3.4

The Scylla team is pleased to announce the release of Scylla Monitoring Stack 3.4

Scylla Monitoring Stack is an open-source stack for monitoring Scylla Enterprise and Scylla Open Source, based on Prometheus and Grafana. Scylla Monitoring Stack 3.4 supports:

  • Scylla Open Source versions 3.2, 3.3, 4.0 and the upcoming 4.1
  • Scylla Enterprise versions 2018.x and 2019.x
  • Scylla Manager 2.0.x and 2.1.x

Related Links

New in Scylla Monitoring Stack 3.4

  • Dashboards for Scylla Open Source 4.1 #926
  • Upgrade to Grafana 6.7.3 from 6.6 #855
  • Upgrade to Prometheus 2.18.1 from 2.15.1 #854
  • Support range annotations #852
    With the addition of Grafana’s range annotation, the Manager task annotations are now shown as a range.This is how a running repair looks like on the graphs.

  • Scylla Manager 2.1 added agent and server system metrics #874
    There is a new Manager agent section in the Manager dashboard that monitors the Manager-Agent that runs on each of the nodes.
  • Set units “ops, rps, wps to ops/s reads/s and writes/s #737
    Some of the dashboards units were unclear, for example, ops can mean both operations/seconds or operations. Using Grafana user-defined types, it was changed to more precise types.
  • Add a quick navigation that keeps the time selection #946
    The filtering line has now a quick dashboard navigation that keeps the time range and variable selection.
  • Scheduling group CPU metrics are now averaged by default #945
    In the CPU dashboard, the usage is now averaged by default with an option to choose another aggregation function.

Operational Changes

  • Allow saving of provisioned dashboards from the UI #802
    It is now possible to save changes you made to a dashboard. You should do so with care. If you want to make persistent changes to your dashboard you should use an external directory for Grafana (see Scylla Monitoring documentation).
  • node_exporter scrap is set to 1m #919
    node_exporter is an agent that reads the OS related metrics from a node (memory, disk usage and network) reducing its scrap interval, reduces some unneeded load from the nodes and is safer in case of a slow response.
  • Update the alertmanager plugin to 0.0.8 #917
    Add Apache Licence to the project #263

Bug Fixes

  • Fix relative path for internal links #952
  • Filter out Scylla Manager successful health check annotation #941

17 Jun 2020