ES_SHARD_CACHE_EVICTION_MEM elasticsearch resource_error ai_generated true

ElasticsearchException: shard request cache eviction triggered memory pressure, current heap usage [85%]

ID: elasticsearch/shard-request-cache-eviction-memory-pressure

Also available as: JSON · Markdown · 中文
83%Fix Rate
88%Confidence
1Evidence
2024-09-01First Seen

Version Compatibility

VersionStatusIntroducedDeprecatedNotes
7.17.0 active — — —
8.11.0 active — — —
8.12.0 active — — —

Root Cause

The shard request cache grows too large under high query load, causing frequent evictions that increase heap pressure and degrade performance.

generic

中文

在高查询负载下，分片请求缓存增长过大，导致频繁逐出，增加堆压力并降低性能。

Official Documentation

https://www.elastic.co/guide/en/elasticsearch/reference/current/shard-request-cache.html

Workarounds

  1. 80% success Reduce cache size by setting `indices.requests.cache.size: 1%` in `elasticsearch.yml` to limit heap usage. 
    Reduce cache size by setting `indices.requests.cache.size: 1%` in `elasticsearch.yml` to limit heap usage.
  2. 85% success Use explicit query caching with `?request_cache=true` only for queries that benefit from it, and set `index.requests.cache.enable: false` on indices with high write volume. 
    Use explicit query caching with `?request_cache=true` only for queries that benefit from it, and set `index.requests.cache.enable: false` on indices with high write volume.
  3. 90% success Monitor cache eviction rates via `GET _nodes/stats/indices/request_cache` and optimize queries to reduce cache churn. 
    Monitor cache eviction rates via `GET _nodes/stats/indices/request_cache` and optimize queries to reduce cache churn.

中文步骤

  1. Reduce cache size by setting `indices.requests.cache.size: 1%` in `elasticsearch.yml` to limit heap usage.
  2. Use explicit query caching with `?request_cache=true` only for queries that benefit from it, and set `index.requests.cache.enable: false` on indices with high write volume.
  3. Monitor cache eviction rates via `GET _nodes/stats/indices/request_cache` and optimize queries to reduce cache churn.

Dead Ends

Common approaches that don't work:

  1. 70% fail

    Larger heap delays evictions but doesn't solve the root cause of excessive caching; cache can grow unbounded and cause GC issues.

  2. 65% fail

    Disabling cache increases query latency for repeated queries, especially in dashboards or reporting scenarios.