Diagram
A timeline showing Cloud Run container instances instantly scaling up with a traffic spike, stepping down to zero instances during idle periods for $0 cost, and spinning up again on a new request, while a Compute Engine group with configured minimum instances never reaches zero.
Scaling To Zero During Inactivity
Scaling to zero means a serverless service can reduce its resource usage to zero instances during periods of inactivity. Serverless autoscaling is instant and does not require fine-tuning rules, so when no traffic arrives no compute instances run and no cost is incurred for idle capacity. This behavior is a defining benefit of serverless computing compared with other autoscaling that keeps minimum capacity running.
Google Cloud serverless services that are inherently elastic and can scale to zero include:
Domain 2 of 3
27 practice questions
Practice this DomainLast updated