Add pseudonymization for event properties to Gitlab.com Product Analytics cluster

Problem

With #415557 (closed) we want to start sending events from the Gitlab.com Frontend to Product Analytics. However, as long as we do not have pseudonymization in place we cannot send additional context with the event (e.g. user_id) since that would violate our usage data agreements.

Desired Outcome

Pseudonymization for PII data is set up on the Product Analytics cluster for Gitlab.com in a way that is at least as strong as it is done for the old snowplow pipeline.

Proposed Solution

  1. Add the Pseudonymization Enrichment to the cluster for Gitlab.com
  2. Configure it to pseudonymize user_id using SHA256 like the current pseudonymization.
Edited by Sebastian Rehm