Loading

Slow search on frozen tier nodes

Search on frozen-tier nodes exceeded the frozen-tier latency threshold for consecutive samples. Frozen searches are typically I/O-bound against searchable snapshots, so spikes might reflect cache misses, concurrent searches, or undersized frozen nodes.

Note

For a complete list of insights, refer to AutoOps insights.

Field Value
Component Elasticsearch
Severity High
Scope Node
Domains performance, search

You can customize these settings to adjust when AutoOps detects this event and presents the insight. Refer to AutoOps event settings for details.

The default customization settings are:

Setting Type Default
Search latency threshold (ms) Integer 10000
Consecutive samples above threshold Integer 5
Tip

Raising these thresholds reduces noise but delays detection. Lowering them triggers the insight sooner but can increase alerts during minor blips.

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.

Search latency on es-data-01 and es-data-02 on frozen-tier nodes stayed above your configured threshold for enough consecutive samples. Peak latency in the latest sample was 420 ms. Indices with high search activity: logs-prod-000045 Review query logs or search slow logs on the affected node, tune expensive queries, and check CPU and heap load on the same node if latency stays high.

Note

AutoOps shows different recommendations depending on how their conditions match your deployment or cluster.

When search stays slow on a frozen-tier node, queries against archived data become unpredictable. Frozen searches read from searchable snapshots and are more I/O-bound than other tiers, so spikes often reflect cache misses or concurrent access. Common causes also include CPU or heap pressure and undersized frozen capacity. If latency remains high, narrow your query scope, validate retention settings, and scale frozen capacity if interactive search on archived data is required.