List of Metrics#

This has the list of metrics which are available for usage. In many systems, only a subset of metrics get stored based on how they would be used.

Metric Tiers#

Importance is about storage, not stability: it says which metrics are worth keeping when capacity is limited, not what we promise about their names. For that, see Stability Guarantees and Deprecation Policy — in short, mz_* names come from Materialize itself and we disclose rather than freeze them.

Metrics are grouped by “metricImportance” levels (mzmon-specific). These levels guide which metrics are prioritized in metric stores which have limited capacity.

Two tiers sit above essential, and only series the chart’s own rules write belong to them. The heartbeat tier is the call-home heartbeat: a fixed record of what an install runs and whether its monitoring stack works. The alerts tier is ALERTS, the state of every pending or firing alert. A destination at alerts receives both, and every wider tier includes both. For what the heartbeat carries, see What the heartbeat sends.

The essential metrics are the set of metrics that are critical and you would always want to have available. These are used in alerting.

The recommended metrics are the set of metrics that are used in dashboards and are generally desirable for troubleshooting.

The extended set of metrics are used for optional/experimental dashboards.

The diagnostic set of metrics are used for in-depth troubleshooting and analysis.

In our materialize-monitoring configuration, we also provide an all min-importance for including absolutely everything. This is recommended if you have cheaper metric storage like our bundled Thanos provider.

Heartbeat series#

  • mzmon:build_info:count
    Written by the chart's recording rules: see Recorded Series.
  • mzmon:cluster_created_timestamp_seconds
    Written by the chart's recording rules: see Recorded Series.
  • mzmon:environment_info
    Written by the chart's recording rules: see Recorded Series.
  • mzmon:heartbeat
    Written by the chart's recording rules: see Recorded Series.
  • mzmon:node_uname_info:count
    Written by the chart's recording rules: see Recorded Series.
  • mzmon:rule_evaluation_failures:rate5m
    Written by the chart's recording rules: see Recorded Series.
  • mzmon:rules:count
    Written by the chart's recording rules: see Recorded Series.

Alert state#

Essential Metrics#

FIXME: Some links for alerts mistakenly point at the common-queries page.