Hi, I’d like to improve the ‘PMM mysql queries’ dashboard time loading especially when I select different time ranges than the default one.
Our PMM-server is deployed from the OVA file.
So far, I’ve modified the file /etc/percona-server.conf.d/mysqld.cnf setting 1G for innodb_buffer_pool_size but I’ve not found a significant improvement.
I was thinking in increasing the memory assigned to Prometheus but the FAQ only explain how to do this if you have a docker installation.
I’ve looking in the file /etc/prometheus.yml and it has the comment “Managed by pmm-managed. DO NOT EDIT.”
Reading that, will my changes to /etc/percona-server.conf.d/mysqld.cnf lost in the next upgrade?.
How can I increase the memory assigned to Prometheus? It can help to improve the dashboard time loading?
A plus question: Again FAQ only explain how to control data retention for PMM in a docker instance. How can I do that in the server deployed from the ova file?
First of all, sorry for being a little slow with my response, and thank you for your comments. I am going to raise a documentation request to see if we can add some clarity in this area. As you mention, we cover docker memory management with the METRICS_MEMORY variable but not for virtual appliance. There may be a reason for this that I am personally not aware of, but if I raise a request then we should get some clarity on that.
Meanwhile…
Can I just check that by ‘PMM MySQL Queries’ dashboard you mean the second one on the attached image, please
