Bazel remote cache UNAVAILABLE
A remote-cache RPC failed because the service was unavailable. gRPC defines UNAVAILABLE as most likely transient, but the code alone does not distinguish transport failure, overload, shutdown, or another service-health problem.
Common causes
- DNS, routing, firewall, or proxy interruption
- TLS trust, certificate, or hostname mismatch
- Cache service or backing storage overload
- Load balancer health checks or connection limits
What to check
- Test endpoint resolution and TLS from the same runner
- Correlate errors with cache latency, request rate, and storage health
- Inspect proxy and load-balancer resets
- Verify whether failures affect reads, writes, or both
Primary sources
Use these references to confirm protocol and version-specific behavior.
Investigate with build-level evidence
Hermetiq correlates Bazel actions, cache behavior, Buildbarn infrastructure, and remote-execution timing so teams can move from a status code to the build and system behavior behind it.
Technical review: July 20, 2026 · Verify flags and behavior against the Bazel and Buildbarn versions deployed in your environment.