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 includes experimental OpenSearch indexing for message content. The configuration key remains [elasticsearch], and the implementation uses the OpenSearch client. REST and IMAP message searches currently use MongoDB; the REST OpenSearch query path is not enabled.

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
  • Live indexing is enabled per user via a Redis set (feature:indexing), or globally with enabledFeatureFlags.indexer

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-compatible cluster accessible from WildDuck instances
  • The indexing worker enabled on at least one instance
  • Users must be added to the feature:indexing Redis set for historical indexing; enabledFeatureFlags.indexer can enable live indexing globally

Enabling indexing does not switch message searches to OpenSearch.

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.