Skip to main content
User guide

Platform health

A check-up of the Elastic cluster: cluster state, shards, disk, ILM, time baseline, permissions, with a verdict per item; the paid "AI read" ties failing items together and says what to fix first.

Platform health checks the Elastic platform carrying the logs. Opening the page runs a check-up (free); "AI read" is paid (the platform-ops copilot).

Platform health

Overview

A one-line verdict (OK / items to watch / failing / some checks unavailable) plus three cards: passed, to handle (failing + watch), unavailable (missing permission or API).

Checks

Checks

Grouped into to handle / passed / unavailable, one row each: verdict, what was checked, the observed value, why. Areas covered:

CategoryExamples
ClusterHealth colour, node count, master election
Indices and shardsUnassigned shards, shards per node, oversized / tiny shards, read-only indices
StorageDisk watermarks (low / high / flood), snapshot repositories
LifecycleILM execution errors, log indices without ILM
Time baseline"Future" data in indices, ingest lag; Smart query's time-window hint uses the same data
PermissionsWhat ES_USER lacks (listing indices, writing .rst_copilot_*, reading alerts)
KibanaReachable, version match

"Unavailable" is not an error: without monitor the cluster-level checks cannot run; grant it and run again.

AI read

Paid capability (platform-ops copilot), Standard (14-day trial included) and above. The check-up itself is free.

With non-passing items an "AI read" button appears. The model looks at the items together: which share a root cause, what to fix first and how; it consults the cluster runbooks in Runbooks ("used N knowledge-base passages").

AI read

On this page