Skip to main content

Events Manifest Populator


Event Manifest Populator is an Apache Spark job allowing you to backpopulate a Snowplow event manifest in DynamoDB with the metadata of some or all enriched events from your archive in S3.

This one-off job solves the "cold start" problem for identifying cross-batch natural deduplicates in Snowplow's Relational Database Shredder step.
In other words, without running this job you still will be able to deduplicate events across batches, but if Relational Database Shredder encounters duplicate of event that was shredded before you enabled cross-batch deduplication it will land into shredded/good.


In order to use Event Manifest Populator, you need to have boto2 installed:

$ pip install boto

As next step you need to grab file with instructions to run job on AWS EMR.
You can do it by downloading it directly from Github:

$ wget

Now you can run Event Manifest Populator with a single command (inside a directory with


Task has three required arguments:

  1. Path to enriched events archive. It can be found in aws.s3.buckets.enriched.archive setting in your config.yml.
  2. Local path to Duplicate storage configuration JSON.
  3. Local path to Iglu resolver configuration JSON.

Optionally, you can also pass following arguments:

  • --since to reduce amount of data to be stored in DynamodDB.
    If this option was passed Manifest Populator will process enriched events only after specified date.
    Input date supports two formats: YYYY-MM-dd and YYYY-MM-dd-HH-mm-ss.
  • --log-path to store EMR job logs on S3. Normally, Manifest Populator does not
    produce any logs or output, but if some error occured you'll be able to
    inspect it in EMR logs stored in this path.
  • --profile to specify AWS profile to create this EMR job.
  • --jar to specify S3 path to custom JAR

Duplicate storage configuration JSONโ€‹

The configuration JSON should conform to the amazon_dynamodb_config JSON Schema.

The properties of the schema are:

  1. name: a descriptive name for this Snowplow storage target
  2. accessKeyId: AWS Access Key Id
  3. secretAccessKey: AWS Secret Access Key
  4. awsRegion: AWS region
  5. dynamodbTable: DynamoDB table to store information about processed events
  6. purpose: common for all targets. Amazon DynamoDB supports only "DUPLICATE_TRACKING"
  7. id: (optional) machine-readable config id

Note that Event Manifest Populator must be used only with run ids produced with version of snowplow newer than R73 Cuban Macaw as format of TSV files has been changed.

Was this page helpful?