Disk watermarks wrong configuration
One or more disk watermark thresholds are misconfigured, including possibly the frozen flood-stage setting. Inconsistent or invalid thresholds can trigger unnecessary shard moves, put indices into read-only mode too early or too late, or leave frozen nodes without appropriate disk-pressure protection.
For a complete list of insights, refer to AutoOps insights.
| Field | Value |
|---|---|
| Component | Elasticsearch |
| Severity | Medium |
| Scope | Node |
| Domains | configuration, disk-storage, stability |
The following is an example of what you might see when this insight is triggered. Real insights use live data and links from your deployment or cluster.
Current low watermark: 85%.
Current high watermark: 90%.
Current flood-stage watermark: 95%.
Current frozen flood-stage watermark: 95%.
Maximum node disk size: 500 GB.
AutoOps shows different recommendations depending on how their conditions match your deployment or cluster.
Change disk watermark settings
Condition: Shown when any of low, high, flood-stage, or frozen flood-stage watermark config is missing or invalid (frozen flood-stage coverage pending Beats prerequisite elastic/beats#52341).
Update cluster disk watermark settings to match your capacity plan. Use the action below.
PUT _cluster/settings
{
"persistent": {
"cluster.routing.allocation.disk.watermark.low": "85%",
"cluster.routing.allocation.disk.watermark.high": "90%",
"cluster.routing.allocation.disk.watermark.flood_stage": "95%"
}
}
Requires the manage cluster privilege. Requires Elasticsearch 8.0.0 or later. This action changes cluster or index configuration.
Elasticsearch uses three standard watermark thresholds (low, high, flood stage) plus a separate frozen flood-stage threshold when frozen-tier nodes are present. Misconfiguration of any of these thresholds can lead to: - Unnecessary shard relocations that increase cluster load.
- Indices entering read-only mode too early (absolute values larger than a node's disk) or too late (thresholds set too high).
- Frozen nodes missing appropriate disk-pressure guardrails if the frozen flood-stage setting is invalid or absent. Use percentage-based watermarks when nodes have different disk sizes, such as in hot-warm-cold or frozen-tier deployments. Check that absolute-size thresholds do not exceed the smallest node's disk. Ensure the frozen flood-stage setting is present and valid if your cluster includes frozen-tier nodes.