Skip to content

Job level cpu profilling #24108

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
EdwinaZhu wants to merge 3 commits into
base: main
Choose a base branch
from
Draft

Job level cpu profilling #24108

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
50f20c4
feat(observability): job-level tokio duration WIP
Dec 12, 2025
3d481f1
WIP: add job-level cpu profilling
EdwinaZhu Dec 13, 2025
6144e9d
WIP: add modifications
EdwinaZhu Dec 15, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
51 changes: 38 additions & 13 deletions grafana/dashboard/dev/streaming_actors_tokio.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,20 @@
from ..common import *
from . import section

# def _sum_actors_by_mv() -> str:
# return (
# f" sum(sum({metric('stream_actor_count')}) by (fragment_id)"
# f" * on(fragment_id) group_left(materialized_view_id) table_info)"
# f" by (materialized_view_id)"
# )

def _sum_fragment_metric_by_mv(expr: str) -> str:
return (
f"sum(({expr})"
f" * on(fragment_id) group_left(materialized_view_id)"
f" max by (fragment_id, materialized_view_id) ({metric('table_info')}))"
f" by (materialized_view_id)"
)

@section
def _(outer_panels: Panels):
Expand All @@ -14,10 +28,11 @@ def _(outer_panels: Panels):
"This can be used to estimate the CPU usage of an actor",
[
panels.target(
f"sum(rate({metric('stream_actor_poll_duration')}[$__rate_interval])) by (fragment_id)"
f"/ on(fragment_id) sum({metric('stream_actor_count')}) by (fragment_id)"
f"{_sum_fragment_metric_by_mv(f'sum(rate({metric('stream_actor_poll_duration')}[$__rate_interval])) by (fragment_id)')}"
f" / "
f" {_sum_fragment_metric_by_mv(f'sum({metric('stream_actor_count')}) by (fragment_id)')}"
f" / 1000000000",
"fragment {{fragment_id}}",
"job {{materialized_view_id}}",
),
],
),
Expand All @@ -26,31 +41,41 @@ def _(outer_panels: Panels):
"",
[
panels.target(
f"rate({metric('stream_actor_poll_cnt')}[$__rate_interval]) > 0",
"fragment {{fragment_id}}",
),
_sum_fragment_metric_by_mv(f'rate({metric('stream_actor_poll_cnt')}[$__rate_interval]) > 0'),
"job {{materialized_view_id}}",
)
],
),
panels.timeseries_percentage(
"Tokio: Actor Poll Avg Rate Per Poll",
"",
[
panels.target(
f"rate({metric('stream_actor_poll_duration')}[$__rate_interval]) / (rate({metric('stream_actor_poll_cnt')}[$__rate_interval]) > 0) / 1000000000",
"fragment {{fragment_id}}",
),
f"{_sum_fragment_metric_by_mv(f'rate({metric('stream_actor_poll_duration')}[$__rate_interval])')}"
f" / "
f"{_sum_fragment_metric_by_mv(f'(rate({metric('stream_actor_poll_cnt')}[$__rate_interval]) > 0')}"
f" / 1000000000",
"job {{materialized_view_id}}",
)
],
),
panels.timeseries_percentage(
"Tokio: Actor Idle Rate Per Actor",
"Idle time could be due to no data to process, or waiting for async operations like IO",
[
# panels.target(
# f"sum(rate({metric('stream_actor_idle_duration')}[$__rate_interval])) by (fragment_id)"
# f"/ on(fragment_id) sum({metric('stream_actor_count')}) by (fragment_id)"
# f" / 1000000000",
# "fragment {{fragment_id}}",
# ),
panels.target(
f"sum(rate({metric('stream_actor_idle_duration')}[$__rate_interval])) by (fragment_id)"
f"/ on(fragment_id) sum({metric('stream_actor_count')}) by (fragment_id)"
f"{_sum_fragment_metric_by_mv(f'sum(rate({metric('stream_actor_idle_duration')}[$__rate_interval])) by (fragment_id)')}"
f" / "
f"{_sum_fragment_metric_by_mv(f'sum({metric('stream_actor_count')}) by (fragment_id)')}"
f" / 1000000000",
"fragment {{fragment_id}}",
),
"job {{materialized_view_id}}",
)
],
),
panels.timeseries_actor_ops_small(
Expand Down
Loading