ReadMeFeed / Community knowledge
Real questions. Useful conversations. Find the people who know your stack.
Cloud & infrastructure · #gcp
Recovery needs to recreate the working service and its required data after a machine or process is lost. Which artifacts and state need protection, and how should the restore be checked?
The same project needs to behave consistently on developer machines, in CI and after deployment. Which versions, dependencies and configuration should be recorded?
A worked Cloud Run example references an enabled Secret Manager version as an environment variable. The deployer’s account can access the secret. The service uses a dedicated runtime service account, and new instances fail around startup. Which identity and permission should be checked?
Configuration must be available to the application without exposing credentials in source control, logs or browser code. What belongs in the runtime and which access controls matter?
Goal Lower the ongoing cost of running a low‑traffic workload on Google Cloud while keeping the service responsive to occasional requests. Context The Well‑Architected Framework’s cost‑optimization pillar recommends granular control of performance and cost parameters to maximize business value. It is unclear which Cloud Run‑specific setting (for example, CPU
A failure needs to be narrowed down before settings are changed or operations retried. Which evidence best separates application errors from environment and dependency problems?
A fictional service exposes a secret through an environment variable using latest. After rotation, older instances keep working with the previous credential while newly started instances use the new one. The application team expected an immediate, simultaneous update. What deployment approach gives a more predictable result?
An upgrade needs a compatibility check, a tested release and a recovery path. Which changes deserve particular attention before the new version reaches production?
A service’s storage request is denied. In this illustrative case, Policy Troubleshooter reports that the named principal has the permission on the resource. The team has not checked whether the actual call crosses a VPC Service Controls boundary. Does the allowed result settle the whole access question?