How to Optimize Elasticsearch Performance: Tips and Examples

To optimize Elasticsearch performance, focus on efficient indexing, proper shard and replica settings, and use filters instead of queries when possible. Also, monitor resource usage and tune refresh_interval and cache settings to reduce overhead and speed up searches.

Syntax

Key settings and commands to optimize Elasticsearch include:

• index.number_of_shards: Controls how many pieces your data is split into for parallel processing.
• index.number_of_replicas: Sets how many copies of your data exist for fault tolerance and read speed.
• refresh_interval: How often Elasticsearch refreshes the index to make new data searchable.
• query with filters: Filters are cached and faster than full queries.

json

PUT /my_index/_settings
{
  "index": {
    "number_of_shards": 3,
    "number_of_replicas": 1,
    "refresh_interval": "30s"
  }
}

Example

This example shows how to create an index with optimized shard and replica settings, and how to use a filter in a search query for better performance.

json

PUT /products
{
  "settings": {
    "index": {
      "number_of_shards": 2,
      "number_of_replicas": 1,
      "refresh_interval": "60s"
    }
  },
  "mappings": {
    "properties": {
      "name": { "type": "text" },
      "category": { "type": "keyword" },
      "price": { "type": "float" }
    }
  }
}

GET /products/_search
{
  "query": {
    "bool": {
      "filter": [
        { "term": { "category": "electronics" } },
        { "range": { "price": { "lte": 1000 } } }
      ]
    }
  }
}

Output

{
  "hits": {
    "total": { "value": 5, "relation": "eq" },
    "hits": [
      { "_source": { "name": "Smartphone", "category": "electronics", "price": 999 } },
      { "_source": { "name": "Headphones", "category": "electronics", "price": 199 } }
    ]
  }
}

Common Pitfalls

Common mistakes that hurt Elasticsearch performance include:

• Using too many shards for small datasets, which wastes resources.
• Setting refresh_interval too low, causing frequent costly refreshes.
• Running full-text queries instead of filters when filtering data, which is slower.
• Not monitoring cluster health and ignoring slow queries.

json

GET /my_index/_settings

-- This can be optimized to --

PUT /my_index/_settings
{
  "index": {
    "number_of_shards": 3,
    "refresh_interval": "30s"
  }
}

Quick Reference

Summary tips to optimize Elasticsearch performance:

• Use fewer shards for small to medium indexes.
• Set refresh_interval higher during heavy indexing.
• Use filters instead of queries for exact matches.
• Monitor and tune JVM heap size and garbage collection.
• Use doc_values for sorting and aggregations.

Key Takeaways

Set appropriate shard and replica counts based on data size and query load.

Use filters instead of queries for faster, cached searches.

Adjust refresh_interval to balance indexing speed and search freshness.

Monitor cluster health and resource usage regularly.

Avoid over-sharding and excessive refreshes to reduce overhead.