Workers & Tasks

k8s workloads

Cloud Run (and similar) request-based systems are not a good fit for persistent or scheduled processes you want to run in your container: for example to poll a task queue for async jobs or to run a task every hour to send relevant notifications. Coherence provides the ability to run these in a well-configured private GKE cluster with the same access to VPC for Memorystore and the same access to Cloud SQL via a proxy socket as Cloud Run.

If configured for a service, worker processes will run with the same resources and environment config variables/files that Cloud Run workloads use.

If you don't configure workers/tasks in coherence.yml, no GKE cluster or workloads running on it will be created in your cloud project.

Did this page help you?