celeborn/service
Wang, Fei c3d33daabc [CELEBORN-1627] Introduce instance variable for celeborn dashboard to filter metrics
### What changes were proposed in this pull request?

1. add `instanceLabel` in metrics source, prefer `FQDN:port` than `ip:port` even with `celeborn.network.bind.preferIpAddress=false` before
2. add variable  `instance` with  `label_values(metrics_JVMCPUTime_Value, instance)` same as `celeborn-jvm-dashboard.json`
3. add filter `instance=~"${instance}"` for every metrics
4. add missing `legendFormat` for memory file storage metrics expressions

### Why are the changes needed?

There should be too many celeborn instances in production use case, it is better to add filter with instance.

### Does this PR introduce _any_ user-facing change?
Yes. introduce new variable.

But the instance default value is `ALL`, same behavior as before.

### How was this patch tested?

Config: `celeborn.network.bind.preferIpAddress=false`
<img width="1141" alt="image" src="https://github.com/user-attachments/assets/c3161069-790a-4cb2-8654-6d52cf8e5fb0">
<img width="944" alt="image" src="https://github.com/user-attachments/assets/293b8bd4-252a-459c-aa86-5f4aa75eb594">

<img width="939" alt="image" src="https://github.com/user-attachments/assets/1e1b28af-dd71-4c5b-8285-57473a6c9650">

For JVM metrics, before it was ip:port, and now it is FQDN:port.
<img width="947" alt="image" src="https://github.com/user-attachments/assets/fe00762f-605d-4b5e-b0a4-c586bdc0ec1a">

Closes #2777 from turboFei/legend_base.

Authored-by: Wang, Fei <fwang12@ebay.com>
Signed-off-by: mingji <fengmingxiao.fmx@alibaba-inc.com>
2024-10-09 14:47:03 +08:00
..
src [CELEBORN-1627] Introduce instance variable for celeborn dashboard to filter metrics 2024-10-09 14:47:03 +08:00
pom.xml [CELEBORN-1600] Enable check server dependencies 2024-09-20 15:14:56 +08:00