Index settings

Index settings can be of two types: cluster-level settings that affect all indexes in the cluster and index-level settings that affect individual indexes.

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

Cluster-level index settings

There are two types of cluster settings:

• Static cluster-level index settings are settings that you cannot update while the cluster is running. To update a static setting, you must stop the cluster, update the setting, and then restart the cluster.
• Dynamic cluster-level index settings are settings that you can update at any time.

Static cluster-level index settings

OpenSearch supports the following static cluster-level index settings:

• indices.cache.cleanup_interval (Time unit): Schedules a recurring background task that cleans up expired entries from the cache at the specified interval. Default is 1m (1 minute). For more information, see Index request cache.

• indices.requests.cache.size (String): The cache size as a percentage of the heap size (for example, to use 1% of the heap, specify 1%). Default is 1%. For more information, see Index request cache.

Dynamic cluster-level index settings

OpenSearch supports the following dynamic cluster-level index settings:

• action.auto_create_index (Boolean): Automatically creates an index if the index doesn’t already exist. Also applies any index templates that are configured. Default is true.

• action.destructive_requires_name (Boolean): When set to true, you must specify the index name to delete an index. You cannot delete all indexes or use wildcards. Default is true.

• cluster.default.index.refresh_interval (Time unit): Sets the refresh interval when the index.refresh_interval setting is not provided. This setting can be useful when you want to set a default refresh interval across all indexes in a cluster and support the searchIdle setting. You cannot set the interval lower than the cluster.minimum.index.refresh_interval setting.

• cluster.minimum.index.refresh_interval (Time unit): Sets the minimum refresh interval and applies it to all indexes in the cluster. The cluster.default.index.refresh_interval setting should be higher than this setting’s value. If, during index creation, the index.refresh_interval setting is lower than the minimum set, index creation fails.

• cluster.indices.close.enable (Boolean): Enables closing of open indexes in OpenSearch. Default is true.

• indices.recovery.max_bytes_per_sec (String): Limits the total inbound and outbound recovery traffic for each node. This applies to peer recoveries and snapshot recoveries. Default is 40mb. If you set the recovery traffic value to less than or equal to 0mb, rate limiting will be disabled, which causes recovery data to be transferred at the highest possible rate.

• indices.recovery.max_concurrent_file_chunks (Integer): The number of file chunks sent in parallel for each recovery operation. Default is 2.

• indices.recovery.max_concurrent_operations (Integer): The number of operations sent in parallel for each recovery. Default is 1.

• indices.recovery.max_concurrent_remote_store_streams (Integer): The number of streams to the remote repository that can be opened in parallel when recovering a remote store index. Default is 20.

• indices.replication.max_bytes_per_sec (String): Limits the total inbound and outbound replication traffic for each node. If a value is not specified in the configured value the indices.recovery.max_bytes_per_sec setting is used, which defaults to 40 mb. If you set the replication traffic value to less than or equal to 0 mb, rate limiting is disabled, which causes replication data to be transferred at the highest possible rate.

• indices.fielddata.cache.size (String): The maximum size of the field data cache. May be specified as an absolute value (for example, 8GB) or a percentage of the node heap (for example, 50%). This value is static so you must specify it in the opensearch.yml file. If you don’t specify this setting, the maximum size is unlimited. This value should be smaller than the indices.breaker.fielddata.limit. For more information, see Field data circuit breaker.

• indices.query.bool.max_clause_count (Integer): Defines the maximum product of fields and terms that are queryable simultaneously. Before OpenSearch 2.16, a cluster restart was required in order to apply this static setting. Now dynamic, existing search thread pools may use the old static value initially, causing TooManyClauses exceptions. New thread pools use the updated value. Default is 1024.

• cluster.remote_store.index.path.type (String): The path strategy for the data stored in the remote store. This setting is effective only for remote-store-enabled clusters. This setting supports the following values:

  • fixed: Stores the data in path structure <repository_base_path>/<index_uuid>/<shard_id>/.
  • hashed_prefix: Stores the data in path structure hash(<shard-data-idenitifer>)/<repository_base_path>/<index_uuid>/<shard_id>/.
  • hashed_infix: Stores the data in path structure <repository_base_path>/hash(<shard-data-idenitifer>)/<index_uuid>/<shard_id>/. shard-data-idenitifer is characterized by the index_uuid, shard_id, kind of data (translog, segments), and type of data (data, metadata, lock_files). Default is fixed.
• cluster.remote_store.index.path.hash_algorithm (String): The hash function used to derive the hash value when cluster.remote_store.index.path.type is set to hashed_prefix or hashed_infix. This setting is effective only for remote-store-enabled clusters. This setting supports the following values:
  • fnv_1a_base64: Uses the FNV1a hash function and generates a url-safe 20-bit base64-encoded hash value.
  • fnv_1a_composite_1: Uses the FNV1a hash function and generates a custom encoded hash value that scales well with most remote store options. The FNV1a function generates 64-bit value. The custom encoding uses the most significant 6 bits to create a url-safe base64 character and the next 14 bits to create a binary string. Default is fnv_1a_composite_1.

• cluster.remote_store.translog.transfer_timeout (Time unit): Controls the timeout value while uploading translog and checkpoint files during a sync to the remote store. This setting is applicable only for remote-store-enabled clusters. Default is 30s.

• cluster.remote_store.index.segment_metadata.retention.max_count (Integer): Controls the minimum number of metadata files to keep in the segment repository on a remote store. A value below 1 disables the deletion of stale segment metadata files. Default is 10.

• cluster.remote_store.segment.transfer_timeout (Time unit): Controls the maximum amount of time to wait for all new segments to update after refresh to the remote store. If the upload does not complete within a specified amount of time, it throws a SegmentUploadFailedException error. Default is 30m. It has a minimum constraint of 10m.

• cluster.remote_store.translog.path.prefix (String): Controls the fixed path prefix for translog data on a remote-store-enabled cluster. This setting only applies when the cluster.remote_store.index.path.type setting is either HASHED_PREFIX or HASHED_INFIX. Default is an empty string, "".

• cluster.remote_store.segments.path.prefix (String): Controls the fixed path prefix for segment data on a remote-store-enabled cluster. This setting only applies when the cluster.remote_store.index.path.type setting is either HASHED_PREFIX or HASHED_INFIX. Default is an empty string, "".

• cluster.snapshot.shard.path.prefix (String): Controls the fixed path prefix for snapshot shard-level blobs. This setting only applies when the repository shard_path_type setting is either HASHED_PREFIX or HASHED_INFIX. Default is an empty string, "".

• cluster.default_number_of_replicas (Integer): Controls the default number of replicas for indexes in the cluster. The index-level index.number_of_replicas setting defaults to this value if not configured. Default is 1.

• cluster.thread_pool.<fixed-threadpool>.size (Integer): Controls the sizes of both the fixed and resizable queue thread pools. Overrides the defaults provided in opensearch.yml.

• cluster.thread_pool.<scaling-threadpool>.max (Integer): Sets the maximum size of the scaling thread pool. Overrides the default provided in opensearch.yml.

• cluster.thread_pool.<scaling-threadpool>.core (Integer): Specifies the core size of the scaling thread pool. Overrides the default provided in opensearch.yml.

Before tuning thread pool settings dynamically, note that these are expert-level settings that can potentially destabilize your cluster. Modifying thread pool settings applies the same thread pool size to all nodes, so it’s not recommended for clusters with different hardware for the same roles. Similarly, avoid tuning thread pools shared by both data nodes and cluster manager nodes. After making these changes, we recommend monitoring your cluster to ensure that it remains stable and performs as expected.

Index-level index settings

You can specify index settings at index creation. There are two types of index settings:

• Static index-level index settings are settings that you cannot update while the index is open. To update a static setting, you must close the index, update the setting, and then reopen the index.
• Dynamic index-level index settings are settings that you can update at any time.

Specifying a setting when creating an index

When creating an index, you can specify its static or dynamic settings as follows:

PUT /testindex
{
  "settings": {
    "index.number_of_shards": 1,
    "index.number_of_replicas": 2
  }
}

copy

Static index-level index settings

OpenSearch supports the following static index-level index settings:

• index.number_of_shards (Integer): The number of primary shards in the index. Default is 1.

• index.number_of_routing_shards (Integer): The number of routing shards used to split an index.

• index.shard.check_on_startup (Boolean): Whether the index’s shards should be checked for corruption. Available options are false (do not check for corruption), checksum (check for physical corruption), and true (check for both physical and logical corruption). Default is false.

• index.codec (String): Determines how the index’s stored fields are compressed and stored on disk. This setting impacts the size of the index shards and the performance of the index operations.

  Valid values are:

  • default
  • best_compression
  • zstd (OpenSearch 2.9 and later)
  • zstd_no_dict(OpenSearch 2.9 and later)
  • qat_lz4 (OpenSearch 2.14 and later, on supported systems)
  • qat_deflate (OpenSearch 2.14 and later, on supported systems)

For zstd, zstd_no_dict, qat_lz4, and qat_deflate, you can specify the compression level in the index.codec.compression_level setting. For more information, see Index codec settings. Optional. Default is default.

• index.codec.compression_level (Integer): The compression level setting provides a trade-off between compression ratio and speed. A higher compression level results in a higher compression ratio (smaller storage size), but slower compression and decompression speeds lead to higher indexing and search latencies. This setting can only be specified if index.codec is set to zstd and zstd_no_dict in OpenSearch 2.9 and later or qat_lz4 and qat_deflate in OpenSearch 2.14 and later. Valid values are integers in the [1, 6] range. For more information, see Index codec settings. Optional. Default is 3.

• index.codec.qatmode (String): The hardware acceleration mode used for the qat_lz4 and qat_deflate compression codecs. Valid values are auto and hardware. For more information, see Index codec settings. Optional. Default is auto.

• index.routing_partition_size (Integer): The number of shards a custom routing value can go to. Routing helps an imbalanced cluster by relocating values to a subset of shards rather than a single shard. To enable routing, set this value to greater than 1 but less than index.number_of_shards. Default is 1.

• index.soft_deletes.retention_lease.period (Time unit): The maximum amount of time to retain a shard’s history of operations. Default is 12h.

• index.load_fixed_bitset_filters_eagerly (Boolean): Whether OpenSearch should preload cached filters. Available options are true and false. Default is true.

• index.hidden (Boolean): Whether the index should be hidden. Hidden indexes are not returned as part of queries that have wildcards. Available options are true and false. Default is false.

• index.merge.policy (String): This setting controls the merge policy for the Lucene segments. The available options are tiered and log_byte_size. The default is tiered, but for time-series data, such as log events, we recommend that you use the log_byte_size merge policy, which can improve query performance when conducting range queries on the @timestamp field. We recommend that you not change the merge policy of an existing index. Instead, configure this setting when creating a new index.

• index.merge_on_flush.enabled (Boolean): This setting controls Apache Lucene’s merge-on-refresh feature that aims to reduce the number of segments by performing merges on refresh (or in terms of OpenSearch, on flush). Default is true.

• index.merge_on_flush.max_full_flush_merge_wait_time (Time unit): This setting sets the amount of time to wait for merges when index.merge_on_flush.enabled is enabled. Default is 10s.

• index.merge_on_flush.policy (default

  merge-on-flush): This setting controls which merge policy should be used when index.merge_on_flush.enabled is enabled. Default is default.

• index.check_pending_flush.enabled (Boolean): This setting controls the Apache Lucene checkPendingFlushOnUpdate index writer setting, which specifies whether an indexing thread should check for pending flushes on an update in order to flush indexing buffers to disk. Default is true.

• index.use_compound_file (Boolean): This setting controls the Apache Lucene useCompoundFile index writer settings, which specifies whether newly written segment files will be packed into a compound file. Default is true.

Updating a static index setting

You can update a static index setting only on a closed index. The following example demonstrates updating the index codec setting.

First, close an index:

POST /testindex/_close

copy

Then update the settings by sending a request to the _settings endpoint:

PUT /testindex/_settings
{
  "index": {
    "codec": "zstd_no_dict",
    "codec.compression_level": 3
  }
}

copy

Last, reopen the index to enable read and write operations:

POST /testindex/_open

copy

For more information about updating settings, including supported query parameters, see Update settings.

Dynamic index-level index settings

OpenSearch supports the following dynamic index-level index settings:

• index.number_of_replicas (Integer): The number of replica shards each primary shard should have. For example, if you have 4 primary shards and set index.number_of_replicas to 3, the index has 12 replica shards. If not set, defaults to cluster.default_number_of_replicas (which is 1 by default).

• index.auto_expand_replicas (String): Whether the cluster should automatically add replica shards based on the number of data nodes. Specify a lower bound and upper limit (for example, 0–9) or all for the upper limit. For example, if you have 5 data nodes and set index.auto_expand_replicas to 0–3, then the cluster does not automatically add another replica shard. However, if you set this value to 0-all and add 2 more nodes for a total of 7, the cluster will expand to now have 6 replica shards. Default is disabled.

• index.search.idle.after (Time unit): The amount of time a shard should wait for a search or get request until it goes idle. Default is 30s.

• index.search.default_pipeline (String): The name of the search pipeline that is used if no pipeline is explicitly set when searching an index. If a default pipeline is set and the pipeline doesn’t exist, then the index requests fail. Use the pipeline name _none to specify no default search pipeline. For more information, see Default search pipeline.

• index.refresh_interval (Time unit): How often the index should refresh, which publishes its most recent changes and makes them available for searching. Can be set to -1 to disable refreshing. Default is 1s.

• index.max_result_window (Integer): The maximum value of from + size for searches of the index. from is the starting index to search from, and size is the number of results to return. Default is 10000.

• index.max_inner_result_window (Integer): The maximum value of from + size that specifies the number of returned nested search hits and most relevant document aggregated during the query. from is the starting index to search from, and size is the number of top hits to return. Default is 100.

• index.max_rescore_window (Integer): The maximum value of window_size for rescore requests to the index. Rescore requests reorder the index’s documents and return a new score, which can be more precise. Default is the same as index.max_inner_result_window or 10000 by default.

• index.max_docvalue_fields_search (Integer): The maximum number of docvalue_fields allowed in a query. Default is 100.

• index.max_script_fields (Integer): The maximum number of script_fields allowed in a query. Default is 32.

• index.max_ngram_diff (Integer): The maximum difference between min_gram and max_gram values for the NGramTokenizer and NGramTokenFilter. Default is 1.

• index.max_shingle_diff (Integer): The maximum difference between max_shingle_size and min_shingle_size to feed into the shingle token filter. Default is 3.

• index.max_refresh_listeners (Integer): The maximum number of refresh listeners each shard is allowed to have.

• index.analyze.max_token_count (Integer): The maximum number of tokens that can be returned from the _analyze API operation. Default is 10000.

• index.highlight.max_analyzed_offset (Integer): The number of characters a highlight request can analyze. Default is 1000000.

• index.max_terms_count (Integer): The maximum number of terms a terms query can accept. Default is 65536.

• index.max_regex_length (Integer): The maximum character length of regex that can be in a regexp query. Default is 1000.

• index.query.default_field (List): A field or list of fields that OpenSearch uses in queries in case a field isn’t specified in the parameters.

• index.query.max_nested_depth (Integer): The maximum number of nesting levels for nested queries. Default is Integer.MAX_VALUE. Minimum is 1 (single nested query).

• index.requests.cache.enable (Boolean): Enables or disables the index request cache. Default is true. For more information, see Index request cache.

• index.routing.allocation.enable (String): Specifies options for the index’s shard allocation. Available options are all (allow allocation for all shards), primaries (allow allocation only for primary shards), new_primaries (allow allocation only for new primary shards), and none (do not allow allocation). Default is all.

• index.routing.rebalance.enable (String): Enables shard rebalancing for the index. Available options are all (allow rebalancing for all shards), primaries (allow rebalancing only for primary shards), replicas (allow rebalancing only for replicas), and none (do not allow rebalancing). Default is all.

• index.gc_deletes (Time unit): The amount of time to retain a deleted document’s version number. Default is 60s.

• index.default_pipeline (String): The default ingest node pipeline for the index. If the default pipeline is set and the pipeline does not exist, then index requests fail. The pipeline name _none specifies that the index does not have an ingest pipeline.

• index.final_pipeline (String): The final ingest node pipeline for the index. If the final pipeline is set and the pipeline does not exist, then index requests fail. The pipeline name _none specifies that the index does not have an ingest pipeline.

• index.optimize_doc_id_lookup.fuzzy_set.enabled (Boolean): This setting controls whether fuzzy_set should be enabled in order to optimize document ID lookups in index or search calls by using an additional data structure, in this case, the Bloom filter data structure. Enabling this setting improves performance for upsert and search operations that rely on document IDs by creating a new data structure (Bloom filter). The Bloom filter allows for the handling of negative cases (that is, IDs being absent in the existing index) through faster off-heap lookups. Note that creating a Bloom filter requires additional heap usage during indexing time. Default is false.

• index.optimize_doc_id_lookup.fuzzy_set.false_positive_probability (Double): Sets the false-positive probability for the underlying fuzzy_set (that is, the Bloom filter). A lower false-positive probability ensures higher throughput for upsert and get operations but results in increased storage and memory use. Allowed values range between 0.01 and 0.50. Default is 0.20.

Updating a dynamic index setting

You can update a dynamic index setting at any time through the API. For example, to update the refresh interval, use the following request:

PUT /testindex/_settings
{
  "index": {
    "refresh_interval": "2s"
  }
}

copy

For more information about updating settings, including supported query parameters, see Update settings.