[[tune-for-search-speed]] == Tune for search speed [float] === Give memory to the filesystem cache Elasticsearch heavily relies on the filesystem cache in order to make search fast. In general, you should make sure that at least half the available memory goes to the filesystem cache so that Elasticsearch can keep hot regions of the index in physical memory. [float] === Use faster hardware If your search is I/O bound, you should investigate giving more memory to the filesystem cache (see above) or buying faster drives. In particular SSD drives are known to perform better than spinning disks. Always use local storage, remote filesystems such as `NFS` or `SMB` should be avoided. Also beware of virtualized storage such as Amazon's `Elastic Block Storage`. Virtualized storage works very well with Elasticsearch, and it is appealing since it is so fast and simple to set up, but it is also unfortunately inherently slower on an ongoing basis when compared to dedicated local storage. If you put an index on `EBS`, be sure to use provisioned IOPS otherwise operations could be quickly throttled. If your search is CPU-bound, you should investigate buying faster CPUs. [float] === Document modeling Documents should be modeled so that search-time operations are as cheap as possible. In particular, joins should be avoided. <> can make queries several times slower and <> relations can make queries hundreds of times slower. So if the same questions can be answered without joins by denormalizing documents, significant speedups can be expected. [float] === Search as few fields as possible The more fields a <> or <> query targets, the slower it is. A common technique to improve search speed over multiple fields is to copy their values into a single field at index time, and then use this field at search time. This can be automated with the <> directive of mappings without having to change the source of documents. Here is an example of an index containing movies that optimizes queries that search over both the name and the plot of the movie by indexing both values into the `name_and_plot` field. [source,console] -------------------------------------------------- PUT movies { "mappings": { "properties": { "name_and_plot": { "type": "text" }, "name": { "type": "text", "copy_to": "name_and_plot" }, "plot": { "type": "text", "copy_to": "name_and_plot" } } } } -------------------------------------------------- [float] === Pre-index data You should leverage patterns in your queries to optimize the way data is indexed. For instance, if all your documents have a `price` field and most queries run <> aggregations on a fixed list of ranges, you could make this aggregation faster by pre-indexing the ranges into the index and using a <> aggregations. For instance, if documents look like: [source,console] -------------------------------------------------- PUT index/_doc/1 { "designation": "spoon", "price": 13 } -------------------------------------------------- and search requests look like: [source,console] -------------------------------------------------- GET index/_search { "aggs": { "price_ranges": { "range": { "field": "price", "ranges": [ { "to": 10 }, { "from": 10, "to": 100 }, { "from": 100 } ] } } } } -------------------------------------------------- // TEST[continued] Then documents could be enriched by a `price_range` field at index time, which should be mapped as a <>: [source,console] -------------------------------------------------- PUT index { "mappings": { "properties": { "price_range": { "type": "keyword" } } } } PUT index/_doc/1 { "designation": "spoon", "price": 13, "price_range": "10-100" } -------------------------------------------------- And then search requests could aggregate this new field rather than running a `range` aggregation on the `price` field. [source,console] -------------------------------------------------- GET index/_search { "aggs": { "price_ranges": { "terms": { "field": "price_range" } } } } -------------------------------------------------- // TEST[continued] [float] [[map-ids-as-keyword]] === Consider mapping identifiers as `keyword` include::../mapping/types/numeric.asciidoc[tag=map-ids-as-keyword] [float] === Avoid scripts If possible, avoid using <> or <> in searches. Because scripts can't make use of index structures, using scripts in search queries can result in slower search speeds. If you often use scripts to transform indexed data, you can speed up search by making these changes during ingest instead. However, that often means slower index speeds. .*Example* [%collapsible] ==== An index, `my_test_scores`, contains two `long` fields: * `math_score` * `verbal_score` When running searches, users often use a script to sort results by the sum of these two field's values. [source,console] ---- GET /my_test_scores/_search { "query": { "term": { "grad_year": "2020" } }, "sort": [ { "_script": { "type": "number", "script": { "source": "doc['math_score'].value + doc['verbal_score'].value" }, "order": "desc" } } ] } ---- // TEST[s/^/PUT my_test_scores\n/] To speed up search, you can perform this calculation during ingest and index the sum to a field instead. First, <>, `total_score`, to the index. The `total_score` field will contain sum of the `math_score` and `verbal_score` field values. [source,console] ---- PUT /my_test_scores/_mapping { "properties": { "total_score": { "type": "long" } } } ---- // TEST[continued] Next, use an <> containing the <> processor to calculate the sum of `math_score` and `verbal_score` and index it in the `total_score` field. [source,console] ---- PUT _ingest/pipeline/my_test_scores_pipeline { "description": "Calculates the total test score", "processors": [ { "script": { "source": "ctx.total_score = (ctx.math_score + ctx.verbal_score)" } } ] } ---- // TEST[continued] To update existing data, use this pipeline to <> any documents from `my_test_scores` to a new index, `my_test_scores_2`. [source,console] ---- POST /_reindex { "source": { "index": "my_test_scores" }, "dest": { "index": "my_test_scores_2", "pipeline": "my_test_scores_pipeline" } } ---- // TEST[continued] Continue using the pipeline to index any new documents to `my_test_scores_2`. [source,console] ---- POST /my_test_scores_2/_doc/?pipeline=my_test_scores_pipeline { "student": "kimchy", "grad_year": "2020", "math_score": 800, "verbal_score": 800 } ---- // TEST[continued] These changes may slow indexing but allow for faster searches. Users can now sort searches made on `my_test_scores_2` using the `total_score` field instead of using a script. [source,console] ---- GET /my_test_scores_2/_search { "query": { "term": { "grad_year": "2020" } }, "sort": [ { "total_score": { "order": "desc" } } ] } ---- // TEST[continued] //// [source,console] ---- DELETE /_ingest/pipeline/my_test_scores_pipeline ---- // TEST[continued] [source,console-result] ---- { "acknowledged": true } ---- //// ==== We recommend testing and benchmarking any indexing changes before deploying them in production. [float] === Search rounded dates Queries on date fields that use `now` are typically not cacheable since the range that is being matched changes all the time. However switching to a rounded date is often acceptable in terms of user experience, and has the benefit of making better use of the query cache. For instance the below query: [source,console] -------------------------------------------------- PUT index/_doc/1 { "my_date": "2016-05-11T16:30:55.328Z" } GET index/_search { "query": { "constant_score": { "filter": { "range": { "my_date": { "gte": "now-1h", "lte": "now" } } } } } } -------------------------------------------------- could be replaced with the following query: [source,console] -------------------------------------------------- GET index/_search { "query": { "constant_score": { "filter": { "range": { "my_date": { "gte": "now-1h/m", "lte": "now/m" } } } } } } -------------------------------------------------- // TEST[continued] In that case we rounded to the minute, so if the current time is `16:31:29`, the range query will match everything whose value of the `my_date` field is between `15:31:00` and `16:31:59`. And if several users run a query that contains this range in the same minute, the query cache could help speed things up a bit. The longer the interval that is used for rounding, the more the query cache can help, but beware that too aggressive rounding might also hurt user experience. NOTE: It might be tempting to split ranges into a large cacheable part and smaller not cacheable parts in order to be able to leverage the query cache, as shown below: [source,console] -------------------------------------------------- GET index/_search { "query": { "constant_score": { "filter": { "bool": { "should": [ { "range": { "my_date": { "gte": "now-1h", "lte": "now-1h/m" } } }, { "range": { "my_date": { "gt": "now-1h/m", "lt": "now/m" } } }, { "range": { "my_date": { "gte": "now/m", "lte": "now" } } } ] } } } } } -------------------------------------------------- // TEST[continued] However such practice might make the query run slower in some cases since the overhead introduced by the `bool` query may defeat the savings from better leveraging the query cache. [float] === Force-merge read-only indices Indices that are read-only may benefit from being <>. This is typically the case with time-based indices: only the index for the current time frame is getting new documents while older indices are read-only. Shards that have been force-merged into a single segment can use simpler and more efficient data structures to perform searches. IMPORTANT: Do not force-merge indices to which you are still writing, or to which you will write again in the future. Instead, rely on the automatic background merge process to perform merges as needed to keep the index running smoothly. If you continue to write to a force-merged index then its performance may become much worse. [float] === Warm up global ordinals Global ordinals are a data-structure that is used in order to run <> aggregations on <> fields. They are loaded lazily in memory because Elasticsearch does not know which fields will be used in `terms` aggregations and which fields won't. You can tell Elasticsearch to load global ordinals eagerly when starting or refreshing a shard by configuring mappings as described below: [source,console] -------------------------------------------------- PUT index { "mappings": { "properties": { "foo": { "type": "keyword", "eager_global_ordinals": true } } } } -------------------------------------------------- [float] === Warm up the filesystem cache If the machine running Elasticsearch is restarted, the filesystem cache will be empty, so it will take some time before the operating system loads hot regions of the index into memory so that search operations are fast. You can explicitly tell the operating system which files should be loaded into memory eagerly depending on the file extension using the <> setting. WARNING: Loading data into the filesystem cache eagerly on too many indices or too many files will make search _slower_ if the filesystem cache is not large enough to hold all the data. Use with caution. [float] === Use index sorting to speed up conjunctions <> can be useful in order to make conjunctions faster at the cost of slightly slower indexing. Read more about it in the <>. [float] [[preference-cache-optimization]] === Use `preference` to optimize cache utilization There are multiple caches that can help with search performance, such as the https://en.wikipedia.org/wiki/Page_cache[filesystem cache], the <> or the <>. Yet all these caches are maintained at the node level, meaning that if you run the same request twice in a row, have 1 <> or more and use https://en.wikipedia.org/wiki/Round-robin_DNS[round-robin], the default routing algorithm, then those two requests will go to different shard copies, preventing node-level caches from helping. Since it is common for users of a search application to run similar requests one after another, for instance in order to analyze a narrower subset of the index, using a preference value that identifies the current user or session could help optimize usage of the caches. [float] === Replicas might help with throughput, but not always In addition to improving resiliency, replicas can help improve throughput. For instance if you have a single-shard index and three nodes, you will need to set the number of replicas to 2 in order to have 3 copies of your shard in total so that all nodes are utilized. Now imagine that you have a 2-shards index and two nodes. In one case, the number of replicas is 0, meaning that each node holds a single shard. In the second case the number of replicas is 1, meaning that each node has two shards. Which setup is going to perform best in terms of search performance? Usually, the setup that has fewer shards per node in total will perform better. The reason for that is that it gives a greater share of the available filesystem cache to each shard, and the filesystem cache is probably Elasticsearch's number 1 performance factor. At the same time, beware that a setup that does not have replicas is subject to failure in case of a single node failure, so there is a trade-off between throughput and availability. So what is the right number of replicas? If you have a cluster that has `num_nodes` nodes, `num_primaries` primary shards _in total_ and if you want to be able to cope with `max_failures` node failures at once at most, then the right number of replicas for you is `max(max_failures, ceil(num_nodes / num_primaries) - 1)`. === Tune your queries with the Profile API You can also analyse how expensive each component of your queries and aggregations are using the {ref}/search-profile.html[Profile API]. This might allow you to tune your queries to be less expensive, resulting in a positive performance result and reduced load. Also note that Profile API payloads can be easily visualised for better readability in the {kibana-ref}/xpack-profiler.html[Search Profiler], which is a Kibana dev tools UI available in all X-Pack licenses, including the free X-Pack Basic license. Some caveats to the Profile API are that: - the Profile API as a debugging tool adds significant overhead to search execution and can also have a very verbose output - given the added overhead, the resulting took times are not reliable indicators of actual took time, but can be used comparatively between clauses for relative timing differences - the Profile API is best for exploring possible reasons behind the most costly clauses of a query but isn't intended for accurately measuring absolute timings of each clause [[faster-phrase-queries]] === Faster phrase queries with `index_phrases` The <> field has an <> option that indexes 2-shingles and is automatically leveraged by query parsers to run phrase queries that don't have a slop. If your use-case involves running lots of phrase queries, this can speed up queries significantly. [[faster-prefix-queries]] === Faster prefix queries with `index_prefixes` The <> field has an <> option that indexes prefixes of all terms and is automatically leveraged by query parsers to run prefix queries. If your use-case involves running lots of prefix queries, this can speed up queries significantly. [[faster-filtering-with-constant-keyword]] === Use <> to speed up filtering There is a general rule that the cost of a filter is mostly a function of the number of matched documents. Imagine that you have an index containing cycles. There are a large number of bicycles and many searches perform a filter on `cycle_type: bicycle`. This very common filter is unfortunately also very costly since it matches most documents. There is a simple way to avoid running this filter: move bicycles to their own index and filter bicycles by searching this index instead of adding a filter to the query. Unfortunately this can make client-side logic tricky, which is where `constant_keyword` helps. By mapping `cycle_type` as a `constant_keyword` with value `bicycle` on the index that contains bicycles, clients can keep running the exact same queries as they used to run on the monolithic index and Elasticsearch will do the right thing on the bicycles index by ignoring filters on `cycle_type` if the value is `bicycle` and returning no hits otherwise. Here is what mappings could look like: [source,console] -------------------------------------------------- PUT bicycles { "mappings": { "properties": { "cycle_type": { "type": "constant_keyword", "value": "bicycle" }, "name": { "type": "text" } } } } PUT other_cycles { "mappings": { "properties": { "cycle_type": { "type": "keyword" }, "name": { "type": "text" } } } } -------------------------------------------------- We are splitting our index in two: one that will contain only bicycles, and another one that contains other cycles: unicycles, tricycles, etc. Then at search time, we need to search both indices, but we don't need to modify queries. [source,console] -------------------------------------------------- GET bicycles,other_cycles/_search { "query": { "bool": { "must": { "match": { "description": "dutch" } }, "filter": { "term": { "cycle_type": "bicycle" } } } } } -------------------------------------------------- // TEST[continued] On the `bicycles` index, Elasticsearch will simply ignore the `cycle_type` filter and rewrite the search request to the one below: [source,console] -------------------------------------------------- GET bicycles,other_cycles/_search { "query": { "match": { "description": "dutch" } } } -------------------------------------------------- // TEST[continued] On the `other_cycles` index, Elasticsearch will quickly figure out that `bicycle` doesn't exist in the terms dictionary of the `cycle_type` field and return a search response with no hits. This is a powerful way of making queries cheaper by putting common values in a dedicated index. This idea can also be combined across multiple fields: for instance if you track the color of each cycle and your `bicycles` index ends up having a majority of black bikes, you could split it into a `bicycles-black` and a `bicycles-other-colors` indices. The `constant_keyword` is not strictly required for this optimization: it is also possible to update the client-side logic in order to route queries to the relevant indices based on filters. However `constant_keyword` makes it transparently and allows to decouple search requests from the index topology in exchange of very little overhead.