o-nikolas opened a new pull request, #40017:
URL: https://github.com/apache/airflow/pull/40017
This change delivers the main portion of hybrid executors to the Airflow
scheduler. Each loop the query for TIs from the DB may contain TIs tagged with
any executor (or no specific executor, which means run on the default
executor). We now sort those TIs by executor after the query and queue tasks
for the executors which have space to take on more TIs.
We now also heartbeat all executors and process their events each scheduler
loop.
This is a second version of these changes which are fairly concise in source
code diff. Most of the PR is testing changes/additions.
## Benchmarking
Here is some benchmarking to show performance of the scheduler with the
changes.
Please refer to the orange (90th quantile) and yellow/green (50th quantile)
plots, blue is the 99th quantile which leads to very spikey data.
Important plots include the bottom two (which are the critical sections of
the main loop), as well as the loop duration as a whole.
### Baseline from main
### Hybrid Scheduler
### Method:
This data was collected while running the below DAG. Leveraging the standard
Airflow metrics along with some metric math (for example the scheduler loop
time minus the executor heartbeat time, middle-left plot), using statsd and
grafana within the same breeze environment for all data collection.
More samples were taken than just the runs included above, but a
representative selection was chosen.
### DAG code:
```
from datetime import datetime
from airflow.models import DAG
from airflow.operators.bash import BashOperator
from airflow.operators.empty import EmptyOperator
initial_scale = 4
max_scale = 5
scaling_factor = 2
dag_id = "for_loop_tasks_80"
with DAG(
dag_id=dag_id,
catchup=False,
schedule_interval="@once",
start_date=datetime(2023, 1, 1),
max_active_runs=1,
) as dag:
start = EmptyOperator(task_id="start")
end = EmptyOperator(task_id="end")
for i in range(80):
task = BashOperator(task_id=f"task_{i}", bash_command="date")
start >> task >> end
```
<!--
Licensed to the Apache Software Foundation (ASF) under one
or more contributor license agreements. See the NOTICE file
distributed with this work for additional information
regarding copyright ownership. The ASF licenses this file
to you under the Apache License, Version 2.0 (the
"License"); you may not use this file except in compliance
with the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing,
software distributed under the License is distributed on an
"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
KIND, either express or implied. See the License for the
specific language governing permissions and limitations
under the License.
-->
<!--
Thank you for contributing! Please make sure that your code changes
are covered with tests. And in case of new features or big changes
remember to adjust the documentation.
Feel free to ping committers for the review!
In case of an existing issue, reference it using one of the following:
closes: #ISSUE
related: #ISSUE
How to write a good git commit message:
http://chris.beams.io/posts/git-commit/
-->
<!-- Please keep an empty line above the dashes. -->
---
**^ Add meaningful description above**
Read the **[Pull Request
Guidelines](https://github.com/apache/airflow/blob/main/contributing-docs/05_pull_requests.rst#pull-request-guidelines)**
for more information.
In case of fundamental code changes, an Airflow Improvement Proposal
([AIP](https://cwiki.apache.org/confluence/display/AIRFLOW/Airflow+Improvement+Proposals))
is needed.
In case of a new dependency, check compliance with the [ASF 3rd Party
License Policy](https://www.apache.org/legal/resolved.html#category-x).
In case of backwards incompatible changes please leave a note in a
newsfragment file, named `{pr_number}.significant.rst` or
`{issue_number}.significant.rst`, in
[newsfragments](https://github.com/apache/airflow/tree/main/newsfragments).
--
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]
