Dear all,
May I know how to config PMM to monitor XtraBD cluster??
thanks.
You will need to install PMM client on each PXC node following the instructions in this link:
[url]Percona Monitoring and Management
And then connect the pmm-client to the PMM Server following these instructions:
[url]Percona Monitoring and Management
Then finally start data collection with these instructions:
[url]Percona Monitoring and Management
You may also refer to this page to know how to configure MySQL for PMM:
[url]Percona Monitoring and Management
Let us know if you encounter any difficulties setting up PMM Server and PMM client.
hi,
Yes. i can successfully monitor Mysql server. however, when i select those PXC graph in PMM. there is no data shown. May I missed anything in order to monitor those cluster-related metrics?
Hi racheung , can you confirm that you have metrics appearing on traditional dashboards, such as MySQL Overview - that you see valid data for your PXC instances? If not, please ensure that the output of
pmm-admin list
shows the in Grafana and if it exists we then display in the drop-down list of the two PXC dashboards. Given a PXC server you generally do not have to do anything additional to get the metrics to display.
Please confirm you’re seeing some MySQL dashboards with valid data for your PXC instance and we’ll continue from there. Thanks
hi Michael,
Thanks for your update. Below is the pmm-admin list output.
SERVICE TYPE NAME LOCAL PORT RUNNING DATA SOURCE OPTIONS
mysql:queries 1.1.1.1 - YES root:@unix(/var/lib/mysql/mysql.sock) query_source=perfschema, query
linux:metrics 1.1.1.1 42000 YES -
mysql:metrics 1.1.1.1 42002 YES root:@unix(/var/lib/mysql/mysql.sock)
proxysql:metrics 1.1.1.1 42004 YES stats:***@tcp(localhost:6032)
However, i have no idea on how to check the Lable_values. Can you let me know how to check it?
Regards,
Hi racheung , ok this is good news that is showing in running state. label_values is a Grafana function:
[url]Prometheus | Grafana documentation
This helps us render a list on the PXC Dashboards for instances that are actually PXC nodes. This works based on the presence of a specific variable named wsrep_cluster_name . You can look in PMM for this information in a few of ways: [LIST=1]
[]PXC/Galera Cluster Overview dashboard → Settings > Templating > $cluster - Edit > Preview of values == is this blank?
[]PXC/Galera Cluster Overview dashboard → Edit any graph → Metric lookup – start typing wsrep_cluster_
[*]:9090/graph - start typing “wsrep” and you should see several options appear
[/LIST] The first operation above validates whether Grafana knows of the instances at all or not.
The second and third help us explore the data in Prometheus to see if we are collecting wsrep information.
Please let me know if any of this information exists, and we’ll take it further. I’m sorry I haven’t been able to give you a full start-to-finish explanation already, this is the first time I’ve seen a mysqld_exporter work only partially so I am hoping to diagnose this more fully and incorporate our findings into the product. Thanks for your patience!
Hi,
Does it mean Grafana not able to detect the instances?
Regards,
I’ve had this problem too… Is there any solution to fix this problem?
I’ve had this problem too… Is there any solution to fix it?
Why this variable is empty at Grafana?
Prometheus does not have any mysql_galera_variables_info parameter.
I think the problem is that mysqld_exporter to get wsrep_cluster_name, uses show global VARIABLES but in my case, I get error from this command.
[HTML]mysql> show global VARIABLES ;
ERROR 1682 (HY000): Native table ‘performance_schema’.‘global_variables’ has the wrong structure[/HTML]
So I must fix this problem and then try again.
mee too™
Hi,
I’m quite new to PMM and have configured monitoring for a 3 node Percona 5.7 cluster (with vertical replicas) but can’t see any galera graphs at all.
Can you see output of show global VARIABLES inside mysql?