OpenSearch

OpenSearch is a search and observability suite for unstructured data. Tenzir can send events to OpenSearch and emulate a OpenSearch-compatible Bulk API endpoint.

When sending data to OpenSearch, Tenzir uses the Bulk API and attempts to maximally batch events for throughput, accumulating multiple events before shipping them within a single API call. You can control batching behavior with the max_content_length and buffer_timeout options.

Use to_opensearch for both OpenSearch and Elasticsearch-compatible Bulk API endpoints. There is no separate to_elasticsearch operator.

For more details, see the documentation for the to_opensearch operator.

Tenzir can also present a OpenSearch-compatible REST API via the accept_opensearch operator.

Examples

Send events to a OpenSearch index

Send an example event to the main index:

from {event: "example"}
to_opensearch "1.2.3.4:9200", action="create", index="main"

Instead of treating the entire event as a document to be indexed by OpenSearch, you can designate a nested record as the document:

from {category: "qux", doc_id: "XXX", event: {foo: "bar"}}
to_opensearch "localhost:9200", id=doc_id, doc=event, action="update", index=category

The above example updates the document with ID XXX with the contents from the nested field event.

Accept data by emulating OpenSearch

Tenzir can act as a drop-in replacement for OpenSearch by accepting data via a Bulk API endpoint. This allows you to point your Logstash or Beats instances to Tenzir instead.

accept_opensearch "0.0.0.0:9200", keep_actions=true
publish "opensearch"

This pipeline accepts data on port 9200 and publishes all received events on the opensearch topic for further processing by other pipelines.

Use accept_opensearch for new pipelines that receive bulk ingestion data.

Setting keep_actions=true causes action objects to remain in the stream, like this:

{create:{_index:"filebeat-8.17.3"}} // 👈 action object
{"@timestamp":2025-03-31T13:42:28.068Z,log:{offset:1,file:{path:"/mounted/logfile"}},message:"hello",input:{type:"log"},host:{name:"eb21"},agent:{id:"682cfcf4-f251-4576-abcb-6c8bcadfda08",name:"eb21",type:"filebeat",version:"8.17.3",ephemeral_id:"17f74f6e-36f0-4045-93e6-c549874716df"},ecs:{version:"8.0.0"}}
{create:{_index:"filebeat-8.17.3"}} // 👈 action object
{"@timestamp":2025-03-31T13:42:28.068Z,log:{offset:7,file:{path:"/mounted/logfile"}},message:"this",input:{type:"log"},host:{name:"eb21"},agent:{id:"682cfcf4-f251-4576-abcb-6c8bcadfda08",name:"eb21",type:"filebeat",version:"8.17.3",ephemeral_id:"17f74f6e-36f0-4045-93e6-c549874716df"},ecs:{version:"8.0.0"}}

Ship data via Filebeat

Configure Filebeat as follows to ship data to Tenzir:

output.elasticsearch:
  hosts: ["localhost:9200"]

Set hosts to the endpoint of the Tenzir pipeline accepting data.

Ship data via Logstash

Configure Logstash with the

opensearch output plugin to ship data to Tenzir:

output {
  opensearch {
    hosts => "https://localhost:9200"
  }
}

Set hosts to the endpoint of the Tenzir pipeline accepting data.