@greptileai the one issue you raised was outdated, this PR is ready to merge please update the score.

Hi @vincentkoc, have u tested this plugin e2e with a real backend (like Jaeger or a standard OTel collector) to ensure data exported as expected?
I have tried to fix all these issues, but still don't see any otel data in my backend.

Hi @vincentkoc, have u tested this plugin e2e with a real backend (like Jaeger or a standard OTel collector) to ensure data exported as expected? I have tried to fix all these issues, but still don't see any otel data in my backend.

@feng-95 was working in my logs but collector is not working (as you point out), i think i know the issue so will switch this branch PR to draft to avoid noise as im testing from branch on a live instance.

Hi @vincentkoc ,

I wanted to check in on the current status of this PR—do you consider it fully ready for review and merge at this stage?

Additionally, since the openclaw repository can sometimes be quite slow with reviewing and merging large PRs, have you considered submitting this to the openclaw-cn community: https://github.com/jiulingyun/openclaw-cn as well? The maintainers there tend to merge faster. Getting it live in openclaw-cn would be an excellent way to validate these new capabilities in the wild, which could provide the proof-of-stability needed to accelerate the review and merge process here in the main openclaw repo.

@feng-95 yes this has been 100% tested by hand and coded by hand with some AI assistance (hence the screenshots), all other PRs are broken. Just rebasing the conflicts and formatting which broke after some merge conflicts. I dont feel comfortable commiting to a fork, your welcomed to sync my branch manually if you wish.

Rebased branch and resolved merge conflicts, ready to merge again.

hi @vincentkoc - amazing, thanks for the patch! tried it, too and works! unfortunately the OTel spans are missing the input / output data.
do you have a way in mind of best adding those? happy to collaborate!

@nimarb that is a seperate issue and will be follow-up PRs, feel free to raise an issue and mention this PR and we can go from there. For the interest of the openclaw team small mergable PRs will help move things along faster.

@nimarb @vincentkoc - Hey, folks! I'm setting up an observability ecosystem for OpenClaw and have been tracking this issue for a while; I'm very interested in getting this working. Just wanted to let you know that I'm open to and interested in working in the follow up tasks for the OpenTelemetry instrumentation.

Hi @vincentkoc, congrats on getting this merged! Thanks for the hard work on this issue.

I’m currently trying to verify this on my end, but I’m running into an issue. After enabling the diagnostics-otel plugin and configuring the reporting endpoint, I’m not seeing any traces or metrics arriving at my server.

Could I be missing something in the setup, or is there a specific configuration format required

Here is the configuration I'm using:

 {
  "diagnostics": {
    "enabled": true,
    "otel": {
      "enabled": true,
      "endpoint": "my_otel_collector",
      "protocol": "http/protobuf",
      "serviceName": "openclaw",
      "traces": true,
      "metrics": true,
      "logs": true,
      "flushIntervalMs": 30000
    }
  },
  "plugins": {
    "entries": {
      "diagnostics-otel": {
        "enabled": true
      }
    }
  }

}

Is there any additional logging I can enable to debug why the export might be failing silently?

Thanks a lot!

Is there any additional logging I can enable to debug why the export might be failing silently?

@feng-95 you can start gateway in debug mode and see the diagnostics for OTEL, secondly disable metrics and logs and just use traces first and see, thirdly it could be networking/endpoint related issue.

Is there any additional logging I can enable to debug why the export might be failing silently?

@feng-95 you can start gateway in debug mode and see the diagnostics for OTEL, secondly disable metrics and logs and just use traces first and see, thirdly it could be networking/endpoint related issue.

It seems that the diagnostics-otel plugin currently uses http/json encoding to report data. However, my otel-collector only supports the Protocol Buffer format. I will go throught with http/json format.

And I just wonder would it be possible to add support for Protocol Buffers or make the encoding format configurable in a future update?