elasticsearch resource_error ai_generated true

SearchContextTooLargeException: search context size exceeds limit

ID: elasticsearch/search-context-too-large

Also available as: JSON · Markdown · 中文

78%Fix Rate

82%Confidence

1Evidence

2024-03-10First Seen

Version Compatibility

Version	Status	Introduced	Deprecated	Notes
elasticsearch 8.11	active	—	—	—
elasticsearch 8.12	active	—	—	—
elasticsearch 7.17	active	—	—	—
opensearch 2.10	active	—	—	—

The search request tries to load too many documents or fields into memory, exceeding the search context memory limit.

generic

搜索请求尝试加载过多文档或字段到内存，超出搜索上下文内存限制。

85% success Reduce the search request size by using smaller from/size or using search_after for pagination. Example: POST /my_index/_search { "size": 100, "search_after": [12345, "abc"] }
```
Reduce the search request size by using smaller from/size or using search_after for pagination. Example: POST /my_index/_search { "size": 100, "search_after": [12345, "abc"] }
```
75% success Increase the search context size limit dynamically: PUT _cluster/settings { "persistent": { "search.max_search_context_size": "10mb" } }
```
Increase the search context size limit dynamically: PUT _cluster/settings { "persistent": { "search.max_search_context_size": "10mb" } }
```
80% success Use source filtering to return only needed fields: POST /my_index/_search { "_source": ["field1", "field2"], "size": 1000 }
```
Use source filtering to return only needed fields: POST /my_index/_search { "_source": ["field1", "field2"], "size": 1000 }
```

通过使用较小的from/size或search_after分页来减少搜索请求大小。

动态增加搜索上下文大小限制。

使用source filtering只返回必要字段。

Common approaches that don't work:

85% fail
This setting controls aggregation buckets, not search context memory; no effect on this error.
50% fail
Search context size limit is separate from heap; heap increase may help indirectly but is not the direct fix.
70% fail
Scroll API also creates search contexts; large size worsens the problem.