Console should not be hard-dependent on Schema Registry availability
Hi team,
We're encountering an issue where Redpanda Console fails to start if the Schema Registry is unavailable. From the logs, it's clear that the service attempts to validate connectivity during startup, and if the Schema Registry returns an error (e.g., 404 or is unreachable), the entire Console process exits with a fatal error:
{"level":"fatal","ts":"2025-07-15T12:01:09.925+0800","msg":"failed to create console service","error":"failed to create kafka svc: failed to verify connectivity to schema registry: response is an error. Status: 404 - 404 page not found\n"}
This creates a hard dependency on Schema Registry, making the Console fragile in environments where Schema Registry may be temporarily unavailable or not used at all.
Expected Behavior:
- Console should still start up and provide partial functionality even if the Schema Registry is down.
- Ideally, it should degrade gracefully and show UI warnings or disable schema-related features dynamically.
Version:
- Redpanda Console v2.8.5
Suggestion:
- Make schema registry integration optional at startup.
- Provide a fallback or retry mechanism, or at minimum, a config option to skip the validation on startup.
Thanks for your consideration!
Hey @cobolbaby , I actually thought we already have a flag for controlling the eager connectivity checks. But it seems we only have one for the KafkaAPI and AdminAPI so far. It is easy and very reasonable to add the same for the schema registry, but we won't backport this into the v2.8 release. I recommend updating to v3.1+, I'll try to add this soon.
Alright, as long as the new version supports it, that gives me a stronger reason to push for a service upgrade.