How to Fix a Broken Pushgateway

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

Original topic: 这个pushgateway坏了怎么修复

| username: TiDBer_7Q5CQdQd

How to fix this broken pushgateway?

| username: xfworld | Original post link

Shrink this node, then clean up the data, and start a new one.

You can use the tiup command:

  1. Shrink
  2. Clean up data
  3. Expand
  4. Check the cluster status
| username: dba远航 | Original post link

Check why the log is corrupted.

| username: Kongdom | Original post link

I suggest checking the startup logs to troubleshoot the issue.
You can also directly replace the node by scaling up and down.

| username: 江湖故人 | Original post link

Post the logs for us to take a look.

| username: TiDBer_7Q5CQdQd | Original post link

I only saw the steps for scaling down TiKV and TiDB, but didn’t see the steps for scaling down the monitoring.

| username: Kongdom | Original post link

:thinking: The operations are all the same commands.

| username: TiDBer_7Q5CQdQd | Original post link

Although its status is currently shown as down, I see that the monitoring function is still working. I’m not sure what it is affecting.

| username: Kongdom | Original post link

Can the Grafana site on port 3000 be opened, and does it have data? :thinking:

| username: tidb菜鸟一只 | Original post link

First, use ps -ef | grep prometheus on the corresponding node to check if the process is running. If it is, try using telnet from the control machine to check if the corresponding port is unreachable.