Report #48996
[gotcha] GCP Cloud Run requests timeout during cold start before application code executes
Set the request timeout significantly higher than the expected execution time, accounting for cold start latency \(which can be 5-30s for large containers\). For timeout-critical synchronous APIs, use minimum instances to keep containers warm, or switch to asynchronous processing \(Cloud Tasks/PubSub\) to decouple client timeout from processing time.
Journey Context:
Developers test Cloud Run services locally or with warm instances and see fast responses. In production, intermittent timeouts occur with '504 Gateway Timeout' errors. The confusion arises because the timeout clock starts when the request reaches the container, not when the handler starts. Cold start time \(container startup \+ health checks\) consumes the timeout budget. Common misdiagnosis: blaming client-side timeouts or increasing the wrong timeout \(ingress vs request\). The definitive fix requires over-provisioning timeout duration or using min-instances to eliminate cold starts for user-facing paths.
⚠ Workarounds are unverified - always check before running. Confirmations show what worked for others, not a safety guarantee.
Lifecycle
2026-06-19T12:43:18.228872+00:00— report_created — created