[Nvida_GPU] Enable TSDB #14262
[Nvida_GPU] Enable TSDB #14262
Conversation
andrewkroh
added
dashboard
|
I think, you may have to consider the list of common dimension fields/subset of them, additionally |
| - name: uuid | ||
| type: keyword | ||
| description: | | ||
| Nvidia GPU uuid |
There was a problem hiding this comment.
| Nvidia GPU uuid | |
| Nvidia GPU UUID |
| - name: instance | ||
| type: keyword | ||
| description: > | ||
| Nvidia GPU instance name |
There was a problem hiding this comment.
Is this prometheus instance address or GPU name? This description is not very clear.
| target_field: gpu.labels.instance | ||
| ignore_missing: true | ||
| - rename: | ||
| field: prometheus.labels.job |
There was a problem hiding this comment.
Could there be more than one jobs in a specific timeslice within a GPU (identified by UUID)? If yes, this must be a dimension
There was a problem hiding this comment.
I don't think there would be multiple job or instance values within a single GPU.
The job and instance field values are associated with the Prometheus endpoint.
For example, from the endpoint I’m currently using:
"instance": "192.168.0.192:9400"
"job": "prometheus"
I'm not sure whether these should be considered dimensions, would appreciate your input on that.
There was a problem hiding this comment.
Thanks for clarifying. It may be Ok to not make it as a dimension
| field: prometheus.labels.pci_bus_id | ||
| target_field: gpu.labels.pci_bus_id | ||
| ignore_missing: true | ||
| - rename: |
There was a problem hiding this comment.
Could there be more than one error codes in a specific timeslice within a GPU (identified by UUID)? If yes, this must be a dimension
There was a problem hiding this comment.
@Linu-Elias , can you add the error code also as a dimension?
Co-authored-by: Agi K Thomas <101976829+agithomas@users.noreply.github.com>
💚 Build Succeeded
History
cc @Linu-Elias |
|
|
Package nvidia_gpu - 0.4.0 containing this change is available at https://epr.elastic.co/package/nvidia_gpu/0.4.0/ |
andrewkroh
added
the
documentation
* Initial commit * field_name change * dashboard * dashboard fix * added screenshots * changelog * changelog * field name changes * build fix * field name changes * fix * dashboard fix * typo fix * resolved comments * dashboard fix * System tests added * changelog * docker image update * tsdb * resolve conflicts * added field * update gpu labels * update gpu labels * changelog * sample events * sample events * fix * sample file fix * update changelog * Update packages/nvidia_gpu/changelog.yml Co-authored-by: Agi K Thomas <101976829+agithomas@users.noreply.github.com> * add err_code dimension true * test update * ecs dimension * ecs dimension --------- Co-authored-by: Agi K Thomas <101976829+agithomas@users.noreply.github.com>
Proposed commit message
Checklist
changelog.ymlfile.Author's Checklist
How to test this PR locally
Related issues
Screenshots