Pub/Sub to Elasticsearch template

The Pub/Sub to Elasticsearch template is a streaming pipeline that reads messages from a Pub/Sub subscription, executes a user-defined function (UDF), and writes them to Elasticsearch as documents. The Dataflow template uses Elasticsearch's data streams feature to store time series data across multiple indices while giving you a single named resource for requests. Data streams are well-suited for logs, metrics, traces, and other continuously generated data stored in Pub/Sub.

The template creates a datastream named logs-gcp.DATASET-NAMESPACE, where:

  • DATASET is the value of the dataset template parameter, or pubsub if not specified.
  • NAMESPACE is the value of the namespace template parameter, or default if not specified.

Pipeline requirements

  • The source Pub/Sub subscription must exist and the messages must be encoded in a valid JSON format.
  • A publicly reachable Elasticsearch host on a Google Cloud instance or on Elastic Cloud with Elasticsearch version 7.0 or above. See Google Cloud Integration for Elastic for more details.
  • A Pub/Sub topic for error output.

Template parameters

Required parameters

  • inputSubscription : Pub/Sub subscription to consume the input from. Name should be in the format of 'projects/your-project-id/subscriptions/your-subscription-name' (Example: projects/your-project-id/subscriptions/your-subscription-name).
  • errorOutputTopic : Pub/Sub output topic for publishing failed records in the format of 'projects/your-project-id/topics/your-topic-name'.
  • connectionUrl : The Elasticsearch URL in the format https://hostname:[port]. If using Elastic Cloud, specify the CloudID. (Example: https://elasticsearch-host:9200).
  • apiKey : The Base64-encoded API key to use for authentication.

Optional parameters

  • dataset : The type of logs sent using Pub/Sub, for which we have an out-of-the-box dashboard. Known log types values are audit, vpcflow and firewall. Default 'pubsub'.
  • namespace : An arbitrary grouping, such as an environment (dev, prod, or qa), a team, or a strategic business unit. Default: 'default'.
  • elasticsearchTemplateVersion : Dataflow Template Version Identifier, usually defined by Google Cloud. Defaults to: 1.0.0.
  • javascriptTextTransformGcsPath : The Cloud Storage URI of the .js file that defines the JavaScript user-defined function (UDF) to use. (Example: gs://my-bucket/my-udfs/my_file.js).
  • javascriptTextTransformFunctionName : The name of the JavaScript user-defined function (UDF) to use. For example, if your JavaScript function code is myTransform(inJson) { /*...do stuff...*/ }, then the function name is myTransform. For sample JavaScript UDFs, see UDF Examples (https://github.com/GoogleCloudPlatform/DataflowTemplates#udf-examples).
  • javascriptTextTransformReloadIntervalMinutes : Specifies how frequently to reload the UDF, in minutes. If the value is greater than 0, Dataflow periodically checks the UDF file in Cloud Storage, and reloads the UDF if the file is modified. This parameter allows you to update the UDF while the pipeline is running, without needing to restart the job. If the value is 0, UDF reloading is disabled. The default value is 0.
  • elasticsearchUsername : The Elasticsearch username to authenticate with. If specified, the value of 'apiKey' is ignored.
  • elasticsearchPassword : The Elasticsearch password to authenticate with. If specified, the value of 'apiKey' is ignored.
  • batchSize : The batch size in number of documents. Defaults to: 1000.
  • batchSizeBytes : The batch size in number of bytes. Defaults to: 5242880 (5mb).
  • maxRetryAttempts : The maximum number of retry attempts. Must be greater than zero. Defaults to: no retries.
  • maxRetryDuration : The maximum retry duration in milliseconds. Must be greater than zero. Defaults to: no retries.
  • propertyAsIndex : The property in the document being indexed whose value specifies _index metadata to include with the document in bulk requests. Takes precedence over an _index UDF. Defaults to: none.
  • javaScriptIndexFnGcsPath : The Cloud Storage path to the JavaScript UDF source for a function that specifies _index metadata to include with the document in bulk requests. Defaults to: none.
  • javaScriptIndexFnName : The name of the UDF JavaScript function that specifies _index metadata to include with the document in bulk requests. Defaults to: none.
  • propertyAsId : A property in the document being indexed whose value specifies _id metadata to include with the document in bulk requests. Takes precedence over an _id UDF. Defaults to: none.
  • javaScriptIdFnGcsPath : The Cloud Storage path to the JavaScript UDF source for the function that specifies _id metadata to include with the document in bulk requests. Defaults to: none.
  • javaScriptIdFnName : The name of the UDF JavaScript function that specifies the _id metadata to include with the document in bulk requests. Defaults to: none.
  • javaScriptTypeFnGcsPath : The Cloud Storage path to the JavaScript UDF source for a function that specifies _type metadata to include with documents in bulk requests. Default: none.
  • javaScriptTypeFnName : The name of the UDF JavaScript function that specifies the _type metadata to include with the document in bulk requests. Defaults to: none.
  • javaScriptIsDeleteFnGcsPath : The Cloud Storage path to the JavaScript UDF source for the function that determines whether to delete the document instead of inserting or updating it. The function returns a string value of true or false. Defaults to: none.
  • javaScriptIsDeleteFnName : The name of the UDF JavaScript function that determines whether to delete the document instead of inserting or updating it. The function returns a string value of true or false. Defaults to: none.
  • usePartialUpdate : Whether to use partial updates (update rather than create or index, allowing partial documents) with Elasticsearch requests. Defaults to: false.
  • bulkInsertMethod : Whether to use INDEX (index, allows upserts) or CREATE (create, errors on duplicate _id) with Elasticsearch bulk requests. Defaults to: CREATE.
  • trustSelfSignedCerts : Whether to trust self-signed certificate or not. An Elasticsearch instance installed might have a self-signed certificate, Enable this to True to by-pass the validation on SSL certificate. (default is False).
  • disableCertificateValidation : If 'true', trust the self-signed SSL certificate. An Elasticsearch instance might have a self-signed certificate. To bypass validation for the certificate, set this parameter to 'true'. Default: false.
  • apiKeyKMSEncryptionKey : The Cloud KMS key to decrypt the API key. This parameter must be provided if the apiKeySource is set to KMS. If this parameter is provided, apiKey string should be passed in encrypted. Encrypt parameters using the KMS API encrypt endpoint. The Key should be in the format projects/{gcp_project}/locations/{key_region}/keyRings/{key_ring}/cryptoKeys/{kms_key_name}. See: https://cloud.google.com/kms/docs/reference/rest/v1/projects.locations.keyRings.cryptoKeys/encrypt (Example: projects/your-project-id/locations/global/keyRings/your-keyring/cryptoKeys/your-key-name).
  • apiKeySecretId : Secret Manager secret ID for the apiKey. This parameter should be provided if the apiKeySource is set to SECRET_MANAGER. Should be in the format projects/{project}/secrets/{secret}/versions/{secret_version}. (Example: projects/your-project-id/secrets/your-secret/versions/your-secret-version).
  • apiKeySource : Source of the API key. One of PLAINTEXT, KMS or SECRET_MANAGER. This parameter must be provided if secret manager or KMS is used. If apiKeySource is set to KMS, apiKeyKMSEncryptionKey and encrypted apiKey must be provided. If apiKeySource is set to SECRET_MANAGER, apiKeySecretId must be provided. If apiKeySource is set to PLAINTEXT, apiKey must be provided. Defaults to: PLAINTEXT.

User-defined functions

This template supports user-defined functions (UDFs) at several points in the pipeline, described below. For more information, see Create user-defined functions for Dataflow templates.

Text transform function

Transforms the Pub/Sub message into an Elasticsearch document.

Template parameters:

  • javascriptTextTransformGcsPath: the Cloud Storage URI of the JavaScript file.
  • javascriptTextTransformFunctionName: the name of the JavaScript function.

Function specification:

  • Input: the Pub/Sub message data field, serialized as a JSON string.
  • Output: a stringified JSON document to insert into Elasticsearch.

Index function

Returns the index to which the document belongs.

Template parameters:

  • javaScriptIndexFnGcsPath: the Cloud Storage URI of the JavaScript file.
  • javaScriptIndexFnName: the name of the JavaScript function.

Function specification:

  • Input: the Elasticsearch document, serialized as a JSON string.
  • Output: the value of the document's _index metadata field.

Document ID function

Returns the document ID.

Template parameters:

  • javaScriptIdFnGcsPath: the Cloud Storage URI of the JavaScript file.
  • javaScriptIdFnName: the name of the JavaScript function.

Function specification:

  • Input: the Elasticsearch document, serialized as a JSON string.
  • Output: the value of the document's _id metadata field.

Document deletion function

Specifies whether to delete a document. To use this function, set the bulk insert mode to INDEX and provide a document ID function.

Template parameters:

  • javaScriptIsDeleteFnGcsPath: the Cloud Storage URI of the JavaScript file.
  • javaScriptIsDeleteFnName: the name of the JavaScript function.

Function specification:

  • Input: the Elasticsearch document, serialized as a JSON string.
  • Output: return the string "true" to delete the document, or "false" to upsert the document.

Mapping type function

Returns the document's mapping type.

Template parameters:

  • javaScriptTypeFnGcsPath: the Cloud Storage URI of the JavaScript file.
  • javaScriptTypeFnName: the name of the JavaScript function.

Function specification:

  • Input: the Elasticsearch document, serialized as a JSON string.
  • Output: the value of the document's _type metadata field.

Run the template

Console

  1. Go to the Dataflow Create job from template page.
    2. Go to Create job from template
  2. In the Job name field, enter a unique job name.
  3. Optional: For Regional endpoint, select a value from the drop-down menu. The default region is us-central1.

    For a list of regions where you can run a Dataflow job, see Dataflow locations.

  4. From the Dataflow template drop-down menu, select the Pub/Sub to Elasticsearch template.
  5. In the provided parameter fields, enter your parameter values.
  6. Click Run job.

gcloud

In your shell or terminal, run the template:

gcloud dataflow flex-template run JOB_NAME \
    --project=PROJECT_ID \
    --region=REGION_NAME \
    --template-file-gcs-location=gs://dataflow-templates-REGION_NAME/VERSION/flex/PubSub_to_Elasticsearch_Flex \
    --parameters \
inputSubscription=SUBSCRIPTION_NAME,\
connectionUrl=CONNECTION_URL,\
dataset=DATASET,\
namespace=NAMESPACE,\
apiKey=APIKEY,\
errorOutputTopic=ERROR_OUTPUT_TOPIC

Replace the following:

  • PROJECT_ID: the Google Cloud project ID where you want to run the Dataflow job
  • JOB_NAME: a unique job name of your choice
  • REGION_NAME: the region where you want to deploy your Dataflow job—for example, us-central1
  • VERSION: the version of the template that you want to use

    You can use the following values:

  • ERROR_OUTPUT_TOPIC: your Pub/Sub topic for error output
  • SUBSCRIPTION_NAME: your Pub/Sub subscription name
  • CONNECTION_URL: your Elasticsearch URL
  • DATASET: your log type
  • NAMESPACE: your namespace for dataset
  • APIKEY: your base64 encoded API key for authentication

API

To run the template using the REST API, send an HTTP POST request. For more information on the API and its authorization scopes, see projects.templates.launch.

POST https://dataflow.googleapis.com/v1b3/projects/PROJECT_ID/locations/LOCATION/flexTemplates:launch
{
   "launch_parameter": {
      "jobName": "JOB_NAME",
      "parameters": {
          "inputSubscription": "SUBSCRIPTION_NAME",
          "connectionUrl": "CONNECTION_URL",
          "dataset": "DATASET",
          "namespace": "NAMESPACE",
          "apiKey": "APIKEY",
          "errorOutputTopic": "ERROR_OUTPUT_TOPIC"
      },
      "containerSpecGcsPath": "gs://dataflow-templates-LOCATION/VERSION/flex/PubSub_to_Elasticsearch_Flex",
   }
}

Replace the following:

  • PROJECT_ID: the Google Cloud project ID where you want to run the Dataflow job
  • JOB_NAME: a unique job name of your choice
  • LOCATION: the region where you want to deploy your Dataflow job—for example, us-central1
  • VERSION: the version of the template that you want to use

    You can use the following values:

  • ERROR_OUTPUT_TOPIC: your Pub/Sub topic for error output
  • SUBSCRIPTION_NAME: your Pub/Sub subscription name
  • CONNECTION_URL: your Elasticsearch URL
  • DATASET: your log type
  • NAMESPACE: your namespace for dataset
  • APIKEY: your base64 encoded API key for authentication

What's next