potiuk commented on code in PR #25121: URL: https://github.com/apache/airflow/pull/25121#discussion_r924186856
########## docs/apache-airflow/howto/dynamic-dag-generation.rst: ########## @@ -140,3 +140,74 @@ Each of them can run separately with related configuration .. warning:: Using this practice, pay attention to "late binding" behaviour in Python loops. See `that GitHub discussion <https://github.com/apache/airflow/discussions/21278#discussioncomment-2103559>`_ for more details + + +Optimizing DAG parsing delays during execution +---------------------------------------------- + +Sometimes when you generate a lot of Dynamic DAGs from a single DAG file, it might cause unnecessary delays +when the DAG file is parsed during task execution. This is visible as delay before a task starts +and when it gets executed. Sometimes the delays might be minutes. + +Why is this happening ? + +You might not be aware but when your task is executed just before execution, airflow parses the Python +file the dag comes from. When tasks are executed, parsing time of dynamically generated DAGs +when task are executed can be optimized. This optimization is most effective when the +number of generated DAGs is high. + +The Airflow Scheduler (or DAG Processor) requires loading of a complete DAG file to process all metadata. +However, task execution requires only a single DAG object to execute a task. Knowing this, allows us to +skip the generation of unnecessary DAG objects when task is executed, shortening the parsing time. +Upon evaluation of a DAG file, command line arguments are supplied which we can use to determine which +airflow component performs parsing: + +* Scheduler/DAG Processor args: ``["scheduler"]`` or ``["dag-processor"]`` +* Task execution args: ``["airflow", "tasks", "run", "dag_id", "task_id", ...]`` + +However, depending on the executor used and forking model, those args might be available via ``sys.args`` +or via name of the process running. Airflow either executes tasks via running a new Python interpreter or +sets the name of the process as "airflow task supervisor: {ARGS}". + +Upon iterating over the collection of things to generate DAGs for, you can use these arguments to determine +whether you need to generate all DAG objects (when parsing in the DAG File processor), or to generate only +a single DAG object (when executing the task). + + +.. code-block:: python + :emphasize-lines: 1,2,3,7,8 + + import sys + import ast + + current_dag = None + if len(sys.argv) > 3 and sys.argv[1] == "tasks": + current_dag = sys.argv[3] + else: + try: + PROCTITLE_PREFIX = "airflow task supervisor: " + proctitle = str(setproctitle.getproctitle()) + if proctitle.startswith(PROCTITLE_PREFIX): + args_string = proctitle[len(PROCTITLE_PREFIX) :] + args = ast.literal_eval(args_string) + if len(args) > 3 and args[1] == "tasks": + current_dag = args[3] + except: + pass Review Comment: Yeah. Comments are needed but rather than repeating in the comment what we actually do, it should explain why we are doing it: ``` # task executed by starting a new Python interpreter ``` and ``` # task executed via forked process ```` are better comments in this case :) -- 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]
