etcd and configuration sync
Troubleshoot etcd connectivity and configuration synchronization issues in Apache APISIX.
APISIX relies on etcd to store and distribute configuration. If the data plane cannot connect to etcd or cannot sync configuration, routes, upstreams, and certificates will stop updating.
etcd connection failure
Symptoms
- APISIX nodes report etcd connection errors.
- New routes or configuration changes do not take effect.
Likely causes
- etcd cluster is unavailable or quorum is lost.
- Network partition between APISIX and etcd.
- etcd TLS or authentication is misconfigured.
Checks
[TODO] Add commands to verify etcd cluster health, network reachability, and TLS settings.
Resolution
[TODO] Add steps to restore etcd, update endpoints, rotate certificates, and restart APISIX.
Verification
[TODO] Add a test that creates a route and confirms it appears on all data plane nodes.
Escalation data
- etcd cluster membership and leader status
- APISIX error logs
config.yamletcd section
Configuration not synchronized
Symptoms
- Admin API accepts changes but the data plane does not reflect them.
- Different APISIX nodes return different responses.
Likely causes
- Stale watch connection on a data plane node.
- etcd keyspace is missing or corrupted.
- APISIX node is using a different etcd prefix.
Checks
[TODO] Add commands to inspect the local configuration cache and compare with etcd.
Resolution
[TODO] Add steps to restart the data plane, verify the etcd prefix, and re-sync from backup.
Verification
[TODO] Add a method to confirm all nodes return the same configuration.
Escalation data
- Local
confandconfig.yaml - etcd key count and revision
- Data plane startup logs