OpenSearch output plugin for Fluentd


For details, see

For an example deployment, see Save all logs to OpenSearch.

Example output configurations

    host: opensearch-cluster.default.svc.cluster.local
    port: 9200
    scheme: https
    ssl_verify: false
    ssl_version: TLSv1_2
      timekey: 1m
      timekey_wait: 30s
      timekey_use_utc: true



Send your logs to OpenSearch

application_name (*string, optional)

Specify the application name for the rollover index to be created.

Default: default

buffer (*Buffer, optional)

bulk_message_request_threshold (string, optional)

Configure bulk_message request splitting threshold size. Default value is 20MB. (20 * 1024 * 1024) If you specify this size as negative number, bulk_message request splitting feature will be disabled.

Default: 20MB

catch_transport_exception_on_retry (*bool, optional)

catch_transport_exception_on_retry (default: true)

Default: true

compression_level (string, optional)


custom_headers (string, optional)

This parameter adds additional headers to request. Example: {"token":"secret"}

Default: {}

customize_template (string, optional)

Specify the string and its value to be replaced in form of hash. Can contain multiple key value pair that would be replaced in the specified template_file. This setting only creates template and to add rollover index please check the rollover_index configuration.

data_stream_enable (*bool, optional)

Use @type opensearch_data_stream

data_stream_name (string, optional)

You can specify Opensearch data stream name by this parameter. This parameter is mandatory for opensearch_data_stream.

data_stream_template_name (string, optional)

Specify an existing index template for the data stream. If not present, a new template is created and named after the data stream.

Default: data_stream_name

default_opensearch_version (int, optional)


Default: 1

emit_error_for_missing_id (bool, optional)


Default: false

emit_error_label_event (*bool, optional)

emit_error_label_event (default: true)

Default: true

endpoint (*OpenSearchEndpointCredentials, optional)

AWS Endpoint Credentials

exception_backup (*bool, optional)

Indicates whether to backup chunk when ignore exception occurs.

Default: true

fail_on_detecting_os_version_retry_exceed (*bool, optional)

fail_on_detecting_os_version_retry_exceed (default: true)

Default: true

fail_on_putting_template_retry_exceed (*bool, optional)

Indicates whether to fail when max_retry_putting_template is exceeded. If you have multiple output plugin, you could use this property to do not fail on Fluentd statup.(default: true)

Default: true

flatten_hashes (bool, optional)

flatten_hashes_separator (string, optional)

Flatten separator

host (string, optional)

You can specify OpenSearch host by this parameter.

Default: localhost

hosts (string, optional)

You can specify multiple OpenSearch hosts with separator “,”. If you specify hosts option, host and port options are ignored.

http_backend (string, optional)

With http_backend typhoeus, the opensearch plugin uses typhoeus faraday http backend. Typhoeus can handle HTTP keepalive.

Default: excon

http_backend_excon_nonblock (*bool, optional)


Default: true

id_key (string, optional)

Field on your data to identify the data uniquely

ignore_exceptions (string, optional)

A list of exception that will be ignored - when the exception occurs the chunk will be discarded and the buffer retry mechanism won’t be called. It is possible also to specify classes at higher level in the hierarchy.

include_index_in_url (bool, optional)

With this option set to true, Fluentd manifests the index name in the request URL (rather than in the request body). You can use this option to enforce an URL-based access control.

include_tag_key (bool, optional)

This will add the Fluentd tag in the JSON record.

Default: false

include_timestamp (bool, optional)

Adds a @timestamp field to the log, following all settings logstash_format does, except without the restrictions on index_name. This allows one to log to an alias in OpenSearch and utilize the rollover API.

Default: false

index_date_pattern (*string, optional)

Specify this to override the index date pattern for creating a rollover index.

Default: now/d

index_name (string, optional)

The index name to write events to

Default: fluentd

index_separator (string, optional)


Default: -

log_os_400_reason (bool, optional)


Default: false

logstash_dateformat (string, optional)

Set the Logstash date format.

Default: %Y.%m.%d

logstash_format (bool, optional)

Enable Logstash log format.

Default: false

logstash_prefix (string, optional)

Set the Logstash prefix.

Default: logstash

logstash_prefix_separator (string, optional)

Set the Logstash prefix separator.

Default: -

max_retry_get_os_version (int, optional)


Default: 15

max_retry_putting_template (string, optional)

You can specify times of retry putting template.

Default: 10

parent_key (string, optional)


password (*secret.Secret, optional)

Password for HTTP Basic authentication. Secret

path (string, optional)

Path for HTTP Basic authentication.

pipeline (string, optional)

This param is to set a pipeline ID of your OpenSearch to be added into the request, you can configure ingest node.

port (int, optional)

You can specify OpenSearch port by this parameter.

Default: 9200

prefer_oj_serializer (bool, optional)

With default behavior, OpenSearch client uses Yajl as JSON encoder/decoder. Oj is the alternative high performance JSON encoder/decoder. When this parameter sets as true, OpenSearch client uses Oj as JSON encoder/decoder.

Default: false

reconnect_on_error (bool, optional)

Indicates that the plugin should reset connection on any error (reconnect on next send). By default it will reconnect only on “host unreachable exceptions”. We recommended to set this true in the presence of OpenSearch shield.

Default: false

reload_after (string, optional)

When reload_connections true, this is the integer number of operations after which the plugin will reload the connections. The default value is 10000.

reload_connections (*bool, optional)

You can tune how the OpenSearch-transport host reloading feature works.(default: true)

Default: true

reload_on_failure (bool, optional)

Indicates that the OpenSearch-transport will try to reload the nodes addresses if there is a failure while making the request, this can be useful to quickly remove a dead node from the list of addresses.

Default: false

remove_keys_on_update (string, optional)

Remove keys on update will not update the configured keys in OpenSearch when a record is being updated. This setting only has any effect if the write operation is update or upsert.

remove_keys_on_update_key (string, optional)

This setting allows remove_keys_on_update to be configured with a key in each record, in much the same way as target_index_key works.

request_timeout (string, optional)

You can specify HTTP request timeout.

Default: 5s

resurrect_after (string, optional)

You can set in the OpenSearch-transport how often dead connections from the OpenSearch-transport’s pool will be resurrected.

Default: 60s

retry_tag (string, optional)

This setting allows custom routing of messages in response to bulk request failures. The default behavior is to emit failed records using the same tag that was provided.

routing_key (string, optional)


ca_file (*secret.Secret, optional)

CA certificate

client_cert (*secret.Secret, optional)

Client certificate

client_key (*secret.Secret, optional)

Client certificate key

client_key_pass (*secret.Secret, optional)

Client key password

scheme (string, optional)

Connection scheme

Default: http

selector_class_name (string, optional)


slow_flush_log_threshold (string, optional)

The threshold for chunk flush performance check. Parameter type is float, not time, default: 20.0 (seconds) If chunk flush takes longer time than this threshold, Fluentd logs a warning message and increases the fluentd_output_status_slow_flush_count metric.

sniffer_class_name (string, optional)

The default Sniffer used by the OpenSearch::Transport class works well when Fluentd has a direct connection to all of the OpenSearch servers and can make effective use of the _nodes API. This doesn’t work well when Fluentd must connect through a load balancer or proxy. The sniffer_class_name parameter gives you the ability to provide your own Sniffer class to implement whatever connection reload logic you require. In addition, there is a new Fluent::Plugin::OpenSearchSimpleSniffer class which reuses the hosts given in the configuration, which is typically the hostname of the load balancer or proxy. For example, a configuration like this would cause connections to logging-os to reload every 100 operations:

ssl_verify (*bool, optional)

Skip ssl verification (default: true)

Default: true

ssl_version (string, optional)

If you want to configure SSL/TLS version, you can specify ssl_version parameter. [SSLv23, TLSv1, TLSv1_1, TLSv1_2]

suppress_doc_wrap (bool, optional)

By default, record body is wrapped by ‘doc’. This behavior can not handle update script requests. You can set this to suppress doc wrapping and allow record body to be untouched.

Default: false

suppress_type_name (*bool, optional)

Suppress type name to avoid warnings in OpenSearch

tag_key (string, optional)

This will add the Fluentd tag in the JSON record.

Default: tag

target_index_affinity (bool, optional)


Default: false

target_index_key (string, optional)

Tell this plugin to find the index name to write to in the record under this key in preference to other mechanisms. Key can be specified as path to nested record using dot (’.’) as a separator.

template_file (*secret.Secret, optional)

The path to the file containing the template to install. Secret

template_name (string, optional)

The name of the template to define. If a template by the name given is already present, it will be left unchanged, unless template_overwrite is set, in which case the template will be updated.

template_overwrite (bool, optional)

Always update the template, even if it already exists.

Default: false

templates (string, optional)

Specify index templates in form of hash. Can contain multiple templates.

time_key (string, optional)

By default, when inserting records in Logstash format, @timestamp is dynamically created with the time at log ingestion. If you’d like to use a custom time, include an @timestamp with your record.

time_key_exclude_timestamp (bool, optional)


Default: false

time_key_format (string, optional)

The format of the time stamp field (@timestamp or what you specify with time_key). This parameter only has an effect when logstash_format is true as it only affects the name of the index we write to.

time_parse_error_tag (string, optional)

With logstash_format true, OpenSearch plugin parses timestamp field for generating index name. If the record has invalid timestamp value, this plugin emits an error event to @ERROR label with time_parse_error_tag configured tag.

time_precision (string, optional)

Should the record not include a time_key, define the degree of sub-second time precision to preserve from the time portion of the routed event.

truncate_caches_interval (string, optional)


unrecoverable_error_types (string, optional)

Default unrecoverable_error_types parameter is set up strictly. Because rejected_execution_exception is caused by exceeding OpenSearch’s thread pool capacity. Advanced users can increase its capacity, but normal users should follow default behavior.

unrecoverable_record_types (string, optional)


use_legacy_template (*bool, optional)

Specify wether to use legacy template or not.

Default: true

user (string, optional)

User for HTTP Basic authentication. This plugin will escape required URL encoded characters within %{} placeholders. e.g. %{demo+}

utc_index (*bool, optional)

By default, the records inserted into index logstash-YYMMDD with UTC (Coordinated Universal Time). This option allows to use local time if you describe utc_index to false.

Default: true

validate_client_version (bool, optional)

When you use mismatched OpenSearch server and client libraries, fluent-plugin-opensearch cannot send data into OpenSearch.

Default: false

verify_os_version_at_startup (*bool, optional)

verify_os_version_at_startup (default: true)

Default: true

with_transporter_log (bool, optional)

This is debugging purpose option to enable to obtain transporter layer log.

Default: false

write_operation (string, optional)

The write_operation can be any of: (index,create,update,upsert)

Default: index


access_key_id (*secret.Secret, optional)

AWS access key id. This parameter is required when your agent is not running on EC2 instance with an IAM Role.

assume_role_arn (*secret.Secret, optional)

Typically, you can use AssumeRole for cross-account access or federation.

assume_role_session_name (*secret.Secret, optional)


assume_role_web_identity_token_file (*secret.Secret, optional)


ecs_container_credentials_relative_uri (*secret.Secret, optional)

Set with AWS_CONTAINER_CREDENTIALS_RELATIVE_URI environment variable value

region (string, optional)

AWS region. It should be in form like us-east-1, us-west-2. Default nil, which means try to find from environment variable AWS_REGION.

secret_access_key (*secret.Secret, optional)

AWS secret key. This parameter is required when your agent is not running on EC2 instance with an IAM Role.

sts_credentials_region (*secret.Secret, optional)

By default, the AWS Security Token Service (AWS STS) is available as a global service, and all AWS STS requests go to a single endpoint at AWS recommends using Regional AWS STS endpoints instead of the global endpoint to reduce latency, build in redundancy, and increase session token validity.

url (string, required)

AWS connection url.