How to Uninstall Monitoring Components (alertmanager, grafana, prometheus) Installed via tiup

This topic has been translated from a Chinese forum by GPT and might contain errors.

Original topic: 想要卸载通过tiup安装的监控 (alertmanager、grafana、prometheus)

| username: TiDBer_7Q5CQdQd

[TiDB Usage Environment] Production Environment
[TiDB Version] 7.3
The monitoring data directory was placed on the system disk, and now the system disk is running out of space. So I want to uninstall (alertmanager, grafana, prometheus) and redeploy them, placing the data on the data disk. Historical data can be discarded. Alternatively, if there’s a way to modify the data directory without uninstalling, that would work too.

| username: zhanggame1 | Original post link

Uninstall using scale-in, and uninstall other components in the same way with IP and port, then scale-out.

Example of uninstalling:
tiup cluster scale-in tidb-test --node --node --node

Adding new ones:
vi scale-out.yaml and add

  • host:
  • host:
  • host:

Then tiup cluster scale-out tidb-test scale-out.yaml -p

| username: zhanggame1 | Original post link

If there is not enough space, you can also adjust the retention period of the data.

| username: Kongdom | Original post link

Yes, it is recommended to use the scale-in method to uninstall.

| username: 小龙虾爱大龙虾 | Original post link

The fastest and simplest solution is to scale up or down.

| username: tidb菜鸟一只 | Original post link

First scale-in, then scale-out is the most reasonable approach.

| username: 普罗米修斯 | Original post link

Scaling up and down is quite fast. If you scale up or down to the data disk, the following method can be used to retain historical monitoring data for reference:

| username: wangccsy | Original post link

What information do you want to monitor?

| username: dba远航 | Original post link

First scale-in, then scale-out.

| username: 烂番薯0 | Original post link

Directly scale down.