anmolxlight opened a new pull request, #68547:
URL: https://github.com/apache/airflow/pull/68547
Fixes #68532
## What
Adds a new `/dags/schedule` page that visualizes the **typical (mean /
median) start and end time of each Dag over a 24-hour timeline**, aggregated
from recent successful Dag runs. This complements the existing per-Dag Gantt
view, which only shows a single run at a time. The new view operates at a
different level: it aggregates across all Dags to answer "across my whole
deployment, when during the day does each Dag usually run?"
## Why
Operators running many Dags currently have no at-a-glance way to see how
scheduled load is distributed across the day. The new view makes the whole
schedule visible at once, which helps to:
- spot overlapping heavy windows where many Dags pile up at the same hour,
- find quiet windows for maintenance / deploys / resource-intensive jobs,
- notice drift when a Dag's typical run time shifts over recent runs,
- plan capacity and SLAs across the fleet.
## How
### Backend
- New `/ui/dag_schedule_overview` endpoint (in
`api_fastapi/core_api/routes/ui/dag_schedule_overview.py`) gated by
`requires_access_dag(method="GET", access_entity=DagAccessEntity.RUN)`.
- New `DagScheduleOverviewService`
(`api_fastapi/core_api/services/ui/dag_schedule_overview.py`) that:
- fetches all DagModel rows (so dags with no runs still appear as
zero-statistic rows),
- pulls up to 200 most-recent successful `DagRun`s per dag, optionally
scoped by a `run_after` window,
- computes mean/median of `start_date` and `end_date` mapped to
seconds-of-day in UTC, plus mean/median duration,
- supports `dag_id_pattern` and `dag_display_name_pattern` filters.
- New datamodels `DagScheduleOverviewEntry` +
`DagScheduleOverviewCollectionResponse`
(`api_fastapi/core_api/datamodels/ui/dag_schedule_overview.py`).
- 11 backend tests in
`tests/unit/api_fastapi/core_api/routes/ui/test_dag_schedule_overview.py`:
200/401/403, all-three-dag fixture, morning/evening/no-runs statistics,
FAILED-run exclusion, pattern filters, run_after window edges.
- OpenAPI `_private_ui.yaml` updated; `pnpm codegen` regenerated.
### Frontend
- New page `src/pages/DagsList/ScheduleOverview/` mounted at
`/dags/schedule` with:
- one row per Dag (linkable to the Dag detail page),
- a 24h tick rail at 0/3/6/9/12/15/18/21/24 hours,
- a horizontal bar showing the mean start→end window,
- a hover tooltip with mean and median start / end, mean and median
duration, and the recent-runs date range,
- a client-side dag-id filter input.
- New sidebar nav button `Schedule Overview` (`FiBarChart2`) on the main
nav, alongside `Dags`.
- New i18n keys under `dags.scheduleOverview.*` and
`common.nav.scheduleOverview` (English locale).
## Verification
```bash
# Backend
uv run --project airflow-core pytest
airflow-core/tests/unit/api_fastapi/core_api/routes/ui/test_dag_schedule_overview.py
--with-db-init
# 11 passed
uv run --project airflow-core ruff check + ruff format --check + mypy
# All clean
```
```bash
# Frontend
pnpm lint # eslint --quiet && tsc -p tsconfig.app.json — passes
pnpm prettier --check public/i18n/locales/en/dags.json
public/i18n/locales/en/common.json
# Passes
```
The two pre-existing vitest failures in `useTagFilter.test.tsx` and
`Logs.test.tsx` reproduce on `main` (verified with `git stash`), so they are
unrelated to this PR.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
