Skip to main content
Temporal Frontend is the Temporal API entry point used by Metoro services. Apiserver and other hub services connect to it when they start workflows, signal workflows, query workflow state, poll task queues, and report workflow or activity progress.

​
Resource Profile

Frontend is a stateless request-routing layer in front of the rest of Temporal. It validates and accepts Temporal API traffic, then coordinates with the History and Matching services to make workflow progress. It should not normally be a major capacity driver for Metoro Hub. Leave the default replica and resource settings alone unless Metoro support recommends a change or there is clear, sustained Frontend-specific pressure.

​
Restarts And Availability

Temporal Frontend pods do not store durable workflow state locally. Workflow state is persisted in PostgreSQL through Temporal persistence. Individual Frontend pods can be killed, restarted, or recreated at any time without losing durable workflow state. During a restart, clients may briefly reconnect or retry requests, then continue through another ready Frontend pod.

​
Before Tuning Frontend

If workflows are slow to start, query, signal, or complete, check these signals before changing Frontend resources:
  • PostgreSQL persistence health, latency, connection pressure, and restarts.
  • Temporal History and Matching health.
  • Frontend pod readiness, restarts, CPU throttling, and scheduling events.
  • Temporal task queue backlog and workflow latency in the Metoro Hub - Temporal dashboard.
  • Network path between Metoro services and the Temporal Frontend service.
Tune Frontend only after the signal clearly points to Frontend capacity rather than persistence, another Temporal component, or the client service using Temporal.