elasticsearch resource_error ai_generated partial

ElasticsearchException: [script_cache] eviction triggered, current heap usage [92%], max cache size [10000]

ID: elasticsearch/script-cache-eviction-memory-pressure

Also available as: JSON · Markdown · 中文

75%Fix Rate

82%Confidence

1Evidence

2024-06-10First Seen

Version Compatibility

Version	Status	Introduced	Deprecated	Notes
7.17.14	active	—	—	—
8.11.0	active	—	—	—
8.14.0	active	—	—	—

The script cache is full and heap memory usage is high, causing Elasticsearch to evict compiled scripts to free memory, which degrades performance.

generic

脚本缓存已满且堆内存使用率很高，导致Elasticsearch驱逐已编译的脚本以释放内存，从而降低性能。

80% success Increase the script cache size in elasticsearch.yml: script.cache.max_size: 20000 and restart nodes, or use dynamic setting: PUT _cluster/settings { "transient": { "script.cache.max_size": "20000" } }
```
Increase the script cache size in elasticsearch.yml: script.cache.max_size: 20000 and restart nodes, or use dynamic setting: PUT _cluster/settings { "transient": { "script.cache.max_size": "20000" } }
```
90% success Reduce script variety by parameterizing inline scripts: use stored scripts or parameters instead of generating unique script sources per query, e.g., { "script": { "id": "my_script", "params": { "factor": 2 } } }
```
Reduce script variety by parameterizing inline scripts: use stored scripts or parameters instead of generating unique script sources per query, e.g., { "script": { "id": "my_script", "params": { "factor": 2 } } }
```

在elasticsearch.yml中增加脚本缓存大小：script.cache.max_size: 20000并重启节点，或使用动态设置：PUT _cluster/settings { "transient": { "script.cache.max_size": "20000" } }

通过参数化内联脚本减少脚本种类：使用存储脚本或参数代替每次查询生成唯一脚本源，例如 { "script": { "id": "my_script", "params": { "factor": 2 } } }

Common approaches that don't work:

60% fail
Larger heap may delay eviction but does not fix the root cause of too many unique scripts being compiled; cache still fills up
90% fail
Disabling the cache forces recompilation on every script execution, causing severe performance degradation and potential OOM