GKE Version Alerting

Andi Lamprecht ·2026-05-11· 2 min read· Accepted

GKE Version Alerting

Context and problem statement

Platform Engineering must create a solution for alerting on GKE versions so they can ensure the availability of Platform services for customers.

Decision drivers

• Actionable alerts

  • Easily created
  • Easily understood
  • Easily delivered
  • Actively work to minimize, reduce, or eliminate Alert Fatigue

• Security

• Accessible and easy-to-use GCP-native services

• Maintainability

• Extensibility

Decision Outcome

Platform Engineering will use a combination of Cloud Functions (Gen2), PubSub, and EventArc to send a message to Slack when an event is emitted by GKE that an UpgradeAvailableEvent exists. Platform Engineering shall be responsible for GKE maintenance and availability.

Consequences

• A paved road does not exist for Google Cloud Function (Gen2) which requires some research and effort.
• Alerting is step one of this journey. Actionable alerts are critically important. Reducing Alert Fatigue is critically important.
• The use of this approach can be extended to create visibility into active UpgradeEvents and Security Patch Bulletins for GKE.
• The use of this approach creates a path for extending event-driven architecture for Platform products in the future with a significant focus on the relationship of Cloud Functions, PubSub, and EventArc.
• The use of this approach potentially creates a path for creating a paved path for an opinionated Google Cloud Function (Gen2).
• Reacting to alerts may destabilize sprint priorities for Platform Engineering in favor of ensuring Platform availability for customers.
• Requires updating OPA to allow new resources google_cloudfunctions2_function_iam_member and google_cloudfunctions2_function.

Links

• Configure cluster notifications for third-party services
• Cluster notifications concepts
• Cluster notifications how-to
• Receive and filter cluster notifications
• Terraform google_cloudfunctions2_function