Diagnose the selected instance.

Identify the instance

Use version and status to confirm the selected local or SSH destination and running daemon. Use an explicit instance directory when multiple instances share a host.

accel --instance /srv/accelerator-a version
accel --instance /srv/accelerator-a status

Check forwarding and CDC

Verify the backend address, database credentials, TLS trust and CDC account permissions. Missing CDC credentials permit forwarding but disable caching. CDC interruptions clear affected cache state and return reads to the database.

Review startup and scope

Confirm the configured warm-up interval has elapsed and the requested schema and objects are within read scope. Inspect hits, misses and cache occupancy alongside application traffic. Configuration edits require restart.

Plan upgrades and recovery

Export configuration before upgrades. Package upgrades preserve the instance configuration; restart explicitly after installation. Restore into a valid instance directory and provision referenced TLS files separately. Cache contents and uncommitted queued writes are not backup contents.

Prepare a support case

Provide the product and package versions, engine version, affected command or workload, relevant error records and reproduction steps. Remove credentials and customer data before sharing logs.