Skip to content

Conversation

@coqian
Copy link
Contributor

@coqian coqian commented Oct 21, 2025

Description

The per node metrics at OSS Ray Data dashboard are not displayed as expected.
Because of this code change #55495, the following three metrics were added a filter for operator, which is not supported by per node metrics, and causes empty result.
ray_data_num_tasks_finished_per_node
ray_data_bytes_outputs_of_finished_tasks_per_node
ray_data_blocks_outputs_of_finished_tasks_per_node

@coqian coqian requested a review from a team as a code owner October 21, 2025 22:18
Copy link
Contributor

@gemini-code-assist gemini-code-assist bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Code Review

This pull request addresses an issue with broken Ray Data per-node metrics on the dashboard by removing an unsupported operator filter from the corresponding Prometheus queries. The changes are applied to the OUTPUT_BYTES_BY_NODE_PANEL, BLOCKS_BY_NODE_PANEL, and TASK_THROUGHPUT_BY_NODE_PANEL definitions. Since the underlying per-node metrics do not have an operator label, removing this filter is the correct fix to ensure data is displayed on these panels. The changes are accurate and well-targeted. I have no further suggestions.

@coqian coqian force-pushed the coqian/fix-per-node-metrics branch from 2023495 to f3b637f Compare October 21, 2025 22:45
@ray-gardener ray-gardener bot added data Ray Data-related issues observability Issues related to the Ray Dashboard, Logging, Metrics, Tracing, and/or Profiling labels Oct 22, 2025
@alanwguo alanwguo added the go add ONLY when ready to merge, run all tests label Oct 22, 2025
Copy link
Member

@bveeramani bveeramani left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@coqian

  • Rather than removing the operator filter, should we add operator as a tag to the per-node gauges?
  • Are these metrics still useful if they're not disambiguated by operator?

@coqian
Copy link
Contributor Author

coqian commented Oct 27, 2025

  • Rather than removing the operator filter, should we add operator as a tag to the per-node gauges?
  • Are these metrics still useful if they're not disambiguated by operator?

@bveeramani According to the original PR, these three metrics are designed to be aggregated at the per node level instead of operator level. And they are useful to provide the visualization segmented by node cc @omatthew98

@bveeramani
Copy link
Member

  • Rather than removing the operator filter, should we add operator as a tag to the per-node gauges?
  • Are these metrics still useful if they're not disambiguated by operator?

@bveeramani According to the original PR, these three metrics are designed to be aggregated at the per node level instead of operator level. And they are useful to provide the visualization segmented by node cc @omatthew98

Huh, okay. If this is consistent with the original intent, SGTM

@bveeramani bveeramani merged commit 85ec43c into ray-project:master Oct 27, 2025
7 checks passed
landscapepainter pushed a commit to landscapepainter/ray that referenced this pull request Nov 17, 2025
…pported operator filter (ray-project#57970)

## Description
The per node metrics at OSS Ray Data dashboard are not displayed as
expected.
Because of this code change ray-project#55495, the following three metrics were
added a filter for `operator`, which is [not
supported](https://github.com/ray-project/ray/blob/e51f8039bc6992d37834bcff109a3d340e78fcde/python/ray/data/_internal/stats.py#L448)
by per node metrics, and causes empty result.
ray_data_num_tasks_finished_per_node
ray_data_bytes_outputs_of_finished_tasks_per_node
ray_data_blocks_outputs_of_finished_tasks_per_node

Signed-off-by: cong.qian <[email protected]>
Aydin-ab pushed a commit to Aydin-ab/ray-aydin that referenced this pull request Nov 19, 2025
…pported operator filter (ray-project#57970)

## Description
The per node metrics at OSS Ray Data dashboard are not displayed as
expected.
Because of this code change ray-project#55495, the following three metrics were
added a filter for `operator`, which is [not
supported](https://github.com/ray-project/ray/blob/e51f8039bc6992d37834bcff109a3d340e78fcde/python/ray/data/_internal/stats.py#L448)
by per node metrics, and causes empty result.
ray_data_num_tasks_finished_per_node
ray_data_bytes_outputs_of_finished_tasks_per_node
ray_data_blocks_outputs_of_finished_tasks_per_node

Signed-off-by: cong.qian <[email protected]>
Signed-off-by: Aydin Abiar <[email protected]>
Future-Outlier pushed a commit to Future-Outlier/ray that referenced this pull request Dec 7, 2025
…pported operator filter (ray-project#57970)

## Description
The per node metrics at OSS Ray Data dashboard are not displayed as
expected.
Because of this code change ray-project#55495, the following three metrics were
added a filter for `operator`, which is [not
supported](https://github.com/ray-project/ray/blob/e51f8039bc6992d37834bcff109a3d340e78fcde/python/ray/data/_internal/stats.py#L448)
by per node metrics, and causes empty result.
ray_data_num_tasks_finished_per_node
ray_data_bytes_outputs_of_finished_tasks_per_node
ray_data_blocks_outputs_of_finished_tasks_per_node

Signed-off-by: cong.qian <[email protected]>
Signed-off-by: Future-Outlier <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

data Ray Data-related issues go add ONLY when ready to merge, run all tests observability Issues related to the Ray Dashboard, Logging, Metrics, Tracing, and/or Profiling

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants