Hello. Is there any possibilities to remove Node from monitoring totally. I don’t even want to see historical data about this Node.
For example, I made a typo in node name and want to fix it. I remove current monitoring and reconfigure pmm-admin config. Now in Grafana I can see 2 nodes with correct and incorrect name.
Hello. Can someone please provide an update? I want to completely remove a node that I registered. From the client, is it done with either pmm-admin or pmm-agent or is there something hidden on the server that can completely remove a node. The node does not have any database registered, just the node exporter. Thanks, Jeff.
I’m curious about this too.
Hello guys, at some moment removed nodes (pmm-admin remove mysql node-name) were disappeared without any additional actions from user.
Hi. PMM 2 retains metrics and query analytics data for 30 days by default (that’s configurable). After that period, it disappears completely without any additional actions.
PMM 1.x has
Hi, how about such case:
I used to have node called “apache18”
|
https://XXX.XXX.XXX.117:42000/metrics |
UP | instance=“apache18” |
A new virtual machine was added with same IP (XXX.XXX.XXX.117), and:
now, i have two monitored nodes in prometheus targets and on grafana plots:
|
https://XXX.XXX.XXX.117:42000/metrics |
UP | instance=“apache18” |
|
https://XXX.XXX.XXX.117:42000/metrics |
UP | instance=“newnode” |
Please help to remove obsolete nodes.
Anyone achived this?
Hi, how about such case:
I used to have node called “apache18”
|
https://XXX.XXX.XXX.117:42000/metrics |
UP | instance=“apache18” |
A new virtual machine was added with same IP (XXX.XXX.XXX.117), and:
now, i have two monitored nodes in prometheus targets and on grafana plots:
|
https://XXX.XXX.XXX.117:42000/metrics |
UP | instance=“apache18” |
|
https://XXX.XXX.XXX.117:42000/metrics |
UP | instance=“newnode” |
Please help with removing obsolete nodes.
Hi, how about such case:
I used to have node called “apache18”
|
https://XXX.XXX.XXX.117:42000/metrics |
UP | instance=“apache18” |
A new virtual machine was added with same IP (XXX.XXX.XXX.117), and:
now, i have two monitored nodes in prometheus targets and on grafana plots:
|
https://XXX.XXX.XXX.117:42000/metrics |
UP | instance=“apache18” |
|
https://XXX.XXX.XXX.117:42000/metrics |
UP | instance=“newnode” |
Please help with removing obsolete nodes.
@djp
You can remove the node registration via the API. There’s an order to getting something under monitoring which means you need to work in reverse to completely remove something. The add order is Node –> Agent –> Service and then the reverse for removal: Service(s) –> Agents –> Node. Take a look under pmm-settings–> pmm-inventory and click the Nodes tab as we’ll start with the Node you want to remove all traces of and get it’s Node ID (copy that value), now go to the agent tab and search for that copied value to see if there are any agents still associated with that node and make note of the Agent ID(s), lastly go to the Services tab and search for your Node ID to see if there are any services still registered and make not of the Service ID(s). From there you have all the information you need and can use the procedures in the swagger interface (https://<pmm-serverAddress>/swagger/ to remove the Services, Agents, and Nodes. And for peace of mind, it is high on our list of priorities to revamp the inventory management and allow much of this to happen in the UI.