Elasticsearch best practice: disable dynamic mapping

📅 Published: Apr 4, 2025 • Thomas Queste

Elasticsearch best practice: disable dynamic mapping in your indices, or even reject unmapped fields.

TL;DR

Disable dynamic mapping in your indices for better control
Consider rejecting unmapped fields for stricter validation
Document your mapping explicitly for maintainability
i.e. Use dynamic: false or dynamic: strict based on your requirements

Why Disable Dynamic Mapping?

Elasticsearch’s dynamic mapping automatically adds fields to your index when it encounters new ones. While convenient for quick starts, it can lead to problems in production:

➡️ Fully documented mapping is a must.
= It’s best when you know what indices contain and how the fields are indexed.

Dynamic mapping can cause:

Unexpected field types
Index bloat with unused fields
Inconsistent mapping across environments
Mapping explosion with high cardinality fields

Tips for Better Mapping

✅ Best practice: use dynamic: false and list all fields explicitly

This prevents new fields from being indexed automatically
Fields not in the mapping will still be stored in _source but won’t be searchable
Add index: false to fields you need to store but don’t need to search

⭐ For stricter validation: use dynamic: strict

Rejects documents with unmapped fields completely
Provides stronger schema validation
Prevents unexpected data from entering your index

Typical Usage

Setting dynamic mapping at index creation:

PUT my_index
{
  "mappings": {
    "dynamic": "false",
    "properties": {
      "title": { "type": "text" },
      "created_at": { "type": "date" },
      "description": { 
        "type": "text", 
        "index": false // field mapped but not indexed
      }
    }
  }
}

Improve this post