Why Is Elasticsearch So Slow? Fixing Slow Searches

How to fix it

Try these in order — the most common fix is first.

1. 1. Find the slow part

   Copy steps
   1. Enable the search slow log
   2. Run the query with the profile API or Kibana Search Profiler (timings are relative)
   Did this fix it? Yes No

2. 2. Fix the cluster

   Copy steps
   1. Leave at least half of memory to the filesystem cache
   2. Use local SSDs where possible
   3. Set search.default_search_timeout
   Did this fix it? Yes No

3. 3. Fix the query and mapping

   Copy steps
   1. Denormalize instead of nested or join
   2. Combine fields with copy_to; map IDs as keyword
   3. Round dates like now-1h/m; avoid scripts
   4. Set a preference value so repeat searches hit the same shard copies
   Did this fix it? Yes No

Symptoms

Queries take seconds, timeouts under load, high CPU or I/O on data nodes.

Why it happens

• Not enough RAM left for the filesystem cache
• Scripts in sorting, aggregations or script_score
• nested and parent-child joins
• Querying many fields or very many shards
• Unrounded dates that defeat the query cache
• No search timeout set

This guide also covers

• why is elasticsearch so slow
• elasticsearch slow queries
• elasticsearch search performance

The same problem in other apps

All slow, laggy or resource-hungry guides →

• Meilisearch Indexing Slow (Meilisearch)
• Typesense Out of Memory (OOM) and RAM Sizing (Typesense)
• Algolia Search Slow or High Latency (Algolia)

Other Elasticsearch errors

• Elasticsearch Cluster Health Yellow: Causes and Fixes
• Elasticsearch Cluster Health Red: How to Fix It
• Elasticsearch Unassigned Shards: Why and How to Fix
• Elasticsearch circuit_breaking_exception: Data Too Large
• Elasticsearch High JVM Memory Pressure and Out of Memory
• Elasticsearch Disk Watermark: Index Read-Only (FORBIDDEN/12)
• Elasticsearch Limit of Total Fields [1000] Exceeded
• Elasticsearch Cannot Change Field Mapping: Reindex Guide