- Services: Service Mesh
- Release Date: Jan. 18, 2023
In a microservices architecture, always define timeouts in your services to avoid waiting indefinitely for responses, which wastes resources and may cause other (waiting) services to fail. There are two types of timeouts in Service Mesh: request timeouts and idle timeouts. The maximum time a microservice has to respond to a request before the client closes the HTTP connection is defined as request timeout. It is possible to configure the timeout per route on the virtual service route table and the ingress gateway route table, as well as per request on the virtual deployment. The idle timeout is defined as the period during which no active requests are received. The connection is closed when the idle timeout is reached, which can be configured per request on the virtual deployment.
Virtual deployment, virtual service route table, and ingress gateway route table documentation cover timeout configuration.