Cluster QPS how is it determiend

In a multiple node cluster how is that qps determined… is it for query’s executed on that server alone?  or does it count the query’s replicated from the other systems to it?

End results:
Attempt to determine (in the ballpark) the amount of questions being asked of the applications.  The same holds true on the pmm software… when looking at user stats on mysql I am assuming that is per system (cant figure out how to return stats from all three systems)

Thank you

Hi, 
Keep in mind that on a PXC cluster reads are linear scalable so if you run any tool like Sysbench for read operations it is possible to say that if one node has 10k QPS for reads the cluster (3 nodes with same hardware and configuration) will have around ~30k QPS for reads if you are using a load balancer/proxy such as ProxySQL.
Now for writes, they are not scalable. This is because writes needs to be certified on all nodes and also there is the conflict resolution so for write QPS if you get 10k QPS using Sysbench this most likely what the entire cluster will reach. A bit of information about Sysbench:
https://www.percona.com/blog/2018/03/05/tpcc-like-workload-sysbench-1-0/
Hope it helps!

Vinicius:

Thank you and I understood the scalability part… my app(s) are a mix of read AND writes… I guess maybe a better way of phrasing it is when doing a status on a single node in a cluster does the status include writes from the other nodes in the cluster. or is it the work THAT node has processed by itself (and then sent to the other nodes)

Hi,
Yes, it includes the writes coming from other nodes as well. Note that all nodes have to do the individual processing of the write statements coming from the other nodes (the certification and apply phases). For reads, you can consider only locally. 

Dont imagine there is a means to tell specifically how many reads each node is doing?

Yes, you can use the InnoDB metrics to get this information. You can use PMM:

https://pmmdemo.percona.com/graph/d/mysql-innodb/mysql-innodb-details?orgId=1&refresh=1m&var-interval=$__auto_interval_interval&var-node_name=ps8&var-crop_host=ps8&var-service_name=ps8-mysql&var-version=8.0.19-10&var-crop_service=ps8-mysql&var-environment=All&var-cluster=All&var-replication_set=All&var-database=All

Or the MySQL metrics itself:

show global status like 'Com_%';

In this case specific:

master [localhost:45008] {msandbox} ((none)) > show global status like 'Com_select';
+---------------+-------+
| Variable_name | Value |
+---------------+-------+
| Com_select | 91 |
+---------------+-------+
1 row in set (0.00 sec)