Elasticsearch

Search, index, and manage documents in Elasticsearch — the distributed full-text search and analytics engine. Use it to power keyword and JSON-query searches, ingest documents into indices, retrieve or delete specific records, and run bulk operations across your cluster.

Overview

Operations

Configuration

Outputs

Search (elasticsearch_search)

Index Document (elasticsearch_index)

Get Document (elasticsearch_get)

Delete Document (elasticsearch_delete)

Bulk Operations (elasticsearch_bulk)

List Indices (elasticsearch_list_indices)

Example

[Starter] → [Elasticsearch: Search] → [Agent: summarize the results]

Connect the Starter block to an Elasticsearch block configured with url set to {{ELASTICSEARCH_URL}} and apiKey set to {{ELASTICSEARCH_API_KEY}}. Set the index to products and pass the query as {"match": {"description": "{{starter.input}}"}}. The Agent block receives {{elasticsearch.hits}} and {{elasticsearch.total}} and produces a human-readable summary of the matching products.

Tips

• API Key is optional — if your Elasticsearch cluster is open (e.g. a local dev node), leave apiKey blank. For Elastic Cloud or any secured deployment, generate an API key in Kibana under Stack Management > API keys and store it as a workflow secret referenced via {{ELASTICSEARCH_API_KEY}}.
• Use Bulk Operations for high-throughput ingestion — instead of looping the Index Document operation, build an NDJSON body and send it in one elasticsearch_bulk call. Check the errors output and inspect items to identify any partial failures.
• Query DSL must be valid JSON — the query param is parsed with JSON.parse before being sent, so double-check the JSON syntax. A minimal match-all query is {"match_all": {}}.
• Index Document auto-ID vs explicit ID — omit documentId to let Elasticsearch assign a unique ID (uses POST /_doc); provide documentId to upsert by a known key (uses PUT /_doc/<id>). The result output tells you whether the document was created or updated.