Don't emit first_task_scheduling_delay metric for only-once dags #12835
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Dags with a schedule interval of None, or `@once` don't have a following schedule, so we can't realistically calculate this metric. Additionally, this changes the emitted metric from seconds to milliseconds -- all timers to statsd should be in milliseconds -- this is what Statsd and apps that consume data from there expect. See apache#10629 for more details. This will be a "breaking" change from 1.10.14, where the metric was back-ported to, but was (incorrectly) emitting seconds.
Member
Author
|
Ping @Acehaidrey -- A bug (that thankfully due to the try except wasn't fatal), and also changing the "unit" from seconds to milliseconds. |
kaxil
approved these changes
Dec 5, 2020
github-actions
bot
added
the
full tests needed
|
The PR most likely needs to run full matrix of tests because it modifies parts of the core of Airflow. However, committers might decide to merge it quickly and take the risk. If they don't merge it quickly - please rebase it to the latest master at your convenience, or amend the last commit of the PR, and push it with --force-with-lease. |
XD-DENG
reviewed
Dec 5, 2020
|
|
||
| ordered_tis_by_start_date = [ti for ti in finished_tis if ti.start_date] | ||
| ordered_tis_by_start_date.sort(key=lambda ti: ti.start_date, reverse=False) | ||
| first_start_date = ordered_tis_by_start_date[0].start_date |
Member
There was a problem hiding this comment.
A question not relating to the changes made in this PR: why not we directly have something like
first_start_date = min(ti.start_date for ti in finished_tis)
Contributor
There was a problem hiding this comment.
hey @XD-DENG yes agree you can do that and simplify this :) Feel free to add that refactor!
XD-DENG
approved these changes
Dec 5, 2020
Member
XD-DENG
left a comment
XD-DENG
left a comment
There was a problem hiding this comment.
The changes in this PR themselves look good to me surely.
Contributor
|
thank you team for adding this check and fixing my issue! sorry. for misssing this check. |
kaxil
pushed a commit
that referenced
this pull request
Dec 7, 2020
) Dags with a schedule interval of None, or `@once` don't have a following schedule, so we can't realistically calculate this metric. Additionally, this changes the emitted metric from seconds to milliseconds -- all timers to statsd should be in milliseconds -- this is what Statsd and apps that consume data from there expect. See #10629 for more details. This will be a "breaking" change from 1.10.14, where the metric was back-ported to, but was (incorrectly) emitting seconds. (cherry picked from commit 4a02e0a)
AntonyRileyAtVerto
pushed a commit
to vertoanalytics/incubator-airflow
that referenced
this pull request
Feb 2, 2021
- BugFix: Tasks with ``depends_on_past`` or ``task_concurrency`` are stuck (apache#12663) - Fix issue with empty Resources in executor_config (apache#12633) - Fix: Deprecated config ``force_log_out_after`` was not used (apache#12661) - Fix empty asctime field in JSON formatted logs (apache#10515) - [AIRFLOW-2809] Fix security issue regarding Flask SECRET_KEY (apache#3651) - [AIRFLOW-2884] Fix Flask SECRET_KEY security issue in www_rbac (apache#3729) - [AIRFLOW-2886] Generate random Flask SECRET_KEY in default config (apache#3738) - Add missing comma in setup.py (apache#12790) - Bugfix: Unable to import Airflow plugins on Python 3.8 (apache#12859) - Fix setup.py missing comma in ``setup_requires`` (apache#12880) - Don't emit first_task_scheduling_delay metric for only-once dags (apache#12835) - Update setup.py to get non-conflicting set of dependencies (apache#12636) - Rename ``[scheduler] max_threads`` to ``[scheduler] parsing_processes`` (apache#12605) - Add metric for scheduling delay between first run task & expected start time (apache#9544) - Add new-style 2.0 command names for Airflow 1.10.x (apache#12725) - Add Kubernetes cleanup-pods CLI command for Helm Chart (apache#11802) - Don't let webserver run with dangerous config (apache#12747) - Replace pkg_resources with importlib.metadata to avoid VersionConflict errors (apache#12694) - Clarified information about supported Databases
cfei18
pushed a commit
to cfei18/incubator-airflow
that referenced
this pull request
Mar 5, 2021
…che#12835) Dags with a schedule interval of None, or `@once` don't have a following schedule, so we can't realistically calculate this metric. Additionally, this changes the emitted metric from seconds to milliseconds -- all timers to statsd should be in milliseconds -- this is what Statsd and apps that consume data from there expect. See apache#10629 for more details. This will be a "breaking" change from 1.10.14, where the metric was back-ported to, but was (incorrectly) emitting seconds. (cherry picked from commit 4a02e0a)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Dags with a schedule interval of None, or
@oncedon't have a followingschedule, so we can't realistically calculate this metric.
Additionally, this changes the emitted metric from seconds to
milliseconds -- all timers to statsd should be in milliseconds -- this
is what Statsd and apps that consume data from there expect. See #10629
for more details.
This will be a "breaking" change from 1.10.14, where the metric was
back-ported to, but was (incorrectly) emitting seconds.
^ Add meaningful description above
Read the Pull Request Guidelines for more information.
In case of fundamental code change, Airflow Improvement Proposal (AIP) is needed.
In case of a new dependency, check compliance with the ASF 3rd Party License Policy.
In case of backwards incompatible changes please leave a note in UPDATING.md.