Search settings

OpenSearch supports the following search settings:

• search.max_buckets (Dynamic, integer): The maximum number of aggregation buckets allowed in a single response. Default is 65535.

• search.phase_took_enabled (Dynamic, Boolean): Enables returning phase-level took time values in search responses. Default is false.

• search.allow_expensive_queries (Dynamic, Boolean): Allows or disallows expensive queries. For more information, see Expensive queries.

• search.query_rewriting.enabled (Dynamic, Boolean): Enables query rewriting optimizations that can improve search performance by transforming queries into more efficient forms. When enabled, OpenSearch can automatically optimize certain query patterns, such as merging multiple term queries on the same field into a single terms query. Default is false.

• search.query_rewriting.terms_threshold (Dynamic, integer): Controls the threshold for the number of term queries on the same field that triggers the terms merging rewriter to combine them into a single terms query. For example, if set to 16 (default), when 16 or more term queries target the same field within a Boolean clause, they will be merged into a single terms query for better performance. Minimum is 2. Default is 16.

• search.query.max_query_string_length (Dynamic, integer): The maximum allowed length for query string queries. This setting helps prevent performance issues by rejecting query strings longer than the specified limit. Default is 32000.

• search.default_allow_partial_results (Dynamic, Boolean): A cluster-level setting that allows returning partial search results if a request times out or a shard fails. If a search request contains an allow_partial_search_results parameter, the parameter takes precedence over this setting. Default is true.

• search.cancel_after_time_interval (Dynamic, time unit): A cluster-level setting that sets the default timeout for all search requests at the coordinating node level. After the specified time has been reached, the request is stopped and all associated tasks are canceled. Default is -1 (no timeout).

• search.default_search_timeout (Dynamic, time unit): A cluster-level setting that specifies the maximum amount of time that a search request can run before the request is canceled at the shard-level. If the timeout interval is specified in the search request, that interval takes precedence over the configured setting. Default is -1.

• search.default_keep_alive (Dynamic, time unit): Specifies the default keep alive value for scroll and Point in Time (PIT) searches. Because a request may land on a shard multiple times (for example, during the query and fetch phases), OpenSearch opens a request context that exists for the full duration of the request to ensure consistency of the shard state for each individual shard request. In a standard search, once the fetch phase completes, the request context is closed. For a scroll or a PIT search, OpenSearch keeps the request context open until explicitly closed (or until the keep alive time is reached). A background thread periodically checks all open scroll and PIT contexts and deletes the ones that have exceeded their keep alive timeout. The search.keep_alive_interval setting specifies how frequently the contexts are checked for expiration. The search.default_keep_alive setting is the default deadline for expiration. A scroll or PIT request can explicitly specify the keep alive, which takes precedence over this setting. Default is 5m.

• search.keep_alive_interval (Static, time unit): Determines the interval at which OpenSearch checks for request contexts that have exceeded their keep alive limit. Default is 1m.

• search.max_keep_alive (Dynamic, time unit): Specifies the maximum keep alive value. The max_keep_alive setting is used as a safety check against the other keep_alive settings (for example, default_keep_alive) and request-level keep alive settings (for scroll and PIT contexts). If a request exceeds the max_keep_alive value in either case, the operation will fail. Default is 24h.

• search.low_level_cancellation (Dynamic, Boolean): Enables low-level request cancellation. Lucene’s classic timeout mechanism only checks the time while collecting search results. However, an expensive query, such as wildcard or prefix, can take a long time to expand before starting to collect results. In this case, the query can run for a period of time that is greater than the timeout value. The low-level cancellation mechanism addresses this scenario by timing out not only while collecting search results but also during the query expansion phase or before performing any Lucene operation. Default is true.

• search.max_open_scroll_context (Dynamic, integer): A node-level setting that specifies the maximum number of open scroll contexts for the node. Default is 500.

• search.request_stats_enabled (Dynamic, Boolean): Turns on node-level collection of phase-timing statistics from the perspective of the coordinator node. The request-level statistics keep track of how long (in total) search requests spend in each of the different search phases. You can retrieve these counters using the Nodes Stats API. Default is false.

• search.highlight.term_vector_multi_value (Static, Boolean): Specifies to highlight snippets across values of a multi-valued field. Default is true.

• search.max_aggregation_rewrite_filters (Dynamic, integer): Determines the maximum number of rewrite filters allowed during aggregation. Set this value to 0 to disable the filter rewrite optimization for aggregations. This is an experimental feature and may change or be removed in future versions.

• search.dynamic_pruning.cardinality_aggregation.max_allowed_cardinality (Dynamic, integer): Determines the threshold for applying dynamic pruning in cardinality aggregation. If a field’s cardinality exceeds this threshold, the aggregation reverts to the default method. This is an experimental feature and may change or be removed in future versions.

• search.aggregation.bucket_selection_strategy_factor (Dynamic, integer): Controls the algorithm used to select top buckets in terms aggregations. This factor determines when to use a priority queue (better for small result sets) and when to use quick select (better for large result sets). The strategy is chosen based on the condition size * factor < bucketsInOrd. A factor of 0 always uses priority queue, while higher values favor quick select for larger result sets. Valid values are 0 to 10 (inclusive). Default is 5.

• search.keyword_index_or_doc_values_enabled (Dynamic, Boolean): Determines whether to use the index or doc values when running multi_term queries on keyword fields. Default value is false.

Scripting settings

OpenSearch supports the following scripting settings:

• script.max_size_in_bytes (Dynamic, byte unit): Controls the maximum script byte size allowed. This setting helps prevent memory issues by rejecting scripts larger than this limit. Default is 65536 (64 KB).

• script.cache.max_size (Static, integer): Sets the maximum number of compiled scripts that can be cached in memory. The script cache stores compiled scripts to avoid recompilation overhead for frequently used scripts. When the cache reaches this limit, the least recently used scripts are evicted to make room for new ones. Increasing this value can improve performance for applications that use many different scripts but will consume more memory. Default is 100.

Point in Time settings

For information about PIT settings, see PIT settings.

To learn more about static and dynamic settings, see Configuring OpenSearch.