Best Practices for Monitoring Percona XtraDB Cluster: What Metrics Matter Most?

Hi all,

I’d like to open a discussion about key metrics to monitor in a Percona XtraDB Cluster environment.

In my opinion, the following three metrics are among the most important for cluster health:

SHOW STATUS LIKE ‘wsrep_cluster_size’;
SHOW STATUS LIKE ‘wsrep_local_state_comment’;
SHOW STATUS LIKE ‘wsrep_cluster_status’;

The expected values should be:

wsrep_cluster_size: total number of nodes (e.g., 3)
wsrep_local_state_comment: should be “Synced”
wsrep_cluster_status`: should be “Primary”

Are there any other metrics you would recommend monitoring, particularly from a performance or reliability point of view?

Any insights or suggestions would be much appreciated.

I suggest you check out Percona Monitoring and Management. A 100% open source, and completely free solution to monitoring all things MySQL and PXC related; developed by MySQL Experts, for MySQL DBAs.

1 Like

Hi @matthewb - thank you so much for your response.

Indeed,I’ve installed PMM in my lab environment, but I have to admit I’m a bit overwhelmed by the large number of available metrics. I’d really appreciate some guidance on which performance metrics are the most important to focus on.

Would you recommend any key metrics such like wsrep_flow_control_paused, wsrep_cert_deps_distance, or wsrep_local_recv_queue_avg?

Am I heading in the right direction with these? And how critical is it to monitor deadlocks in a Percona XtraDB Cluster?

Sharing your own experience would be very much appreciated!