Cluster Health Triage
A 5-step recipe with Elasticsearch. We can help you get access.
Recipe · 5 steps
Cluster Health Triage
with Elasticsearch
Collect a full diagnostic picture of an Elasticsearch cluster, cheapest signal first.
- Read the structured cluster health summary. The status colour and the unassigned shard count are usually enough to classify the incident.
- Read the cat health view for the same signal in the tabular form that gets pasted into an incident channel.
- Read the node roster to confirm every expected node is still a member of the cluster, which is the most common explanation for unassigned shards.
- Pull aggregate cluster statistics for the index count, document totals, and store and heap pressure behind a degraded status.
- Read the full cluster state last. It is the most expensive call in the sweep, so it runs only after the cheaper signals have been captured.
What you'll need to have ready
- verbose: Whether the cat indices style views should include column headers.
For your developer
Arazzo workflow elasticsearch:cluster-health-triage, 5 operations: clusterHealth → catHealth → catNodes → clusterStats → clusterState. Specs on APIs.io.