// the find
ahmetb/cloud-run-faq
Unofficial FAQ and everything you've been wondering about Google Cloud Run.
An unofficial FAQ for Google Cloud Run, maintained by ahmetb, organised as short questions that each open with a direct answer before the explanation. It is for developers deciding whether Cloud Run fits a workload, or already running a service and hitting its sharp edges around cold starts, concurrency, storage and VPC access. It is an opinionated index that points to the official docs, and it says so up front.
Most answers lead with a verdict ('Short answer: No', 'Yes', 'Sort of') and then justify it, which suits this kind of document. The numbers are concrete and checkable: default concurrency of 80 with a ceiling of 1000, billing in 0.1-second steps where App Engine Flexible bills per minute, and a 10-second CPU allowance after SIGTERM. The protocol answer draws a hard line (HTTP/1.x, HTTP/2 and gRPC over TLS, nothing raw TCP or UDP, so no Redis) that saves a reader from testing it. The comparisons with Fargate, Lambda and Azure Container Instances name the real differences in billing and scaling model instead of listing features.
The content is stale in places, and no answer carries a date. The last push was June 2025. The registry answer still presents Container Registry as the main path even though Google moved users to Artifact Registry, and the Terraform answer points at the older google_cloud_run_service resource rather than google_cloud_run_v2_service. The Anthos section links to an anchor (#what-is-cloud-run-on-anthos) that does not exist, because the heading is 'What is Cloud Run for Anthos'. The latency answer is a TODO that shipped ('TODO(ahmetb): Write this section'), and the cold-start guidance stays generic, so the question people search for most is the least developed. Nearly every link carries utm_campaign and utm_source parameters, and the repo ships utmize.sh to generate them, which makes the raw Markdown harder to read and copy from.