Skip to main content

Full-Text Search (OpenSearch)

Full-Text Search

Experimental

This feature is experimental and subject to change. The API, configuration, and behavior may change in future releases without notice.

WildDuck optionally supports OpenSearch (or ElasticSearch) as a full-text search backend for message content. When enabled, it provides faster and more capable search than MongoDB's built-in text indexes.

How It Works​

  • New messages are indexed in real-time via MongoDB change streams (the indexer.js worker watches the journal collection)
  • Historical messages can be bulk-indexed using the user-indexing background task
  • Search queries use OpenSearch instead of MongoDB text indexes
  • The feature is gated per-user via a Redis set (feature:indexing)

Configuration​

[elasticsearch]
enabled = false
url = "http://127.0.0.1:9200"
user = "elastic"
pass = "supersecret"
index = "wildduck"

[elasticsearch.indexer]
enabled = false   # enable the indexing worker

Despite the [elasticsearch] configuration key name, the underlying client library is @opensearch-project/opensearch.

At least one WildDuck instance must have elasticsearch.indexer.enabled = true to process the indexing queue.

Requirements​

  • An OpenSearch (or ElasticSearch) cluster accessible from WildDuck instances
  • The indexing worker enabled on at least one instance
  • Users must be added to the feature:indexing Redis set to enable search for their accounts

Background Tasks​

The user-indexing task type handles bulk indexing for a user's existing messages. This is useful when enabling search for an existing account that already has stored messages.