apache / airflow

Apache Airflow - A platform to programmatically author, schedule, and monitor workflows

Apache Airflow

Apache Airflow (or simply Airflow) is a platform to programmatically author, schedule, and monitor workflows.

When workflows are defined as code, they become more maintainable, versionable, testable, and collaborative.

Use Airflow to author workflows as directed acyclic graphs (DAGs) of tasks. The Airflow scheduler executes your tasks on an array of workers while following the specified dependencies. Rich command line utilities make performing complex surgeries on DAGs a snap. The rich user interface makes it easy to visualize pipelines running in production, monitor progress, and troubleshoot issues when needed.

Table of contents

• Project Focus
• Principles
• Requirements
• Getting started
• Installing from PyPI
• Official source code
• Convenience packages
• User Interface
• Semantic versioning
• Version Life Cycle
• Support for Python and Kubernetes versions
• Base OS support for reference Airflow images
• Approach to dependencies of Airflow
• Contributing
• Voting Policy
• Who uses Apache Airflow?
• Who maintains Apache Airflow?
• What goes into the next release?
• Can I use the Apache Airflow logo in my presentation?
• Airflow merchandise
• Links
• Sponsors

Project Focus

Airflow works best with workflows that are mostly static and slowly changing. When the DAG structure is similar from one run to the next, it clarifies the unit of work and continuity. Other similar projects include Luigi, Oozie and Azkaban.

Airflow is commonly used to process data, but has the opinion that tasks should ideally be idempotent (i.e., results of the task will be the same, and will not create duplicated data in a destination system), and should not pass large quantities of data from one task to the next (though tasks can pass metadata using Airflow's XCom feature). For high-volume, data-intensive tasks, a best practice is to delegate to external services specializing in that type of work.

Airflow is not a streaming solution, but it is often used to process real-time data, pulling data off streams in batches.

Principles

• Dynamic: Airflow pipelines are configuration as code (Python), allowing for dynamic pipeline generation. This allows for writing code that instantiates pipelines dynamically.
• Extensible: Easily define your own operators, executors and extend the library so that it fits the level of abstraction that suits your environment.
• Elegant: Airflow pipelines are lean and explicit. Parameterizing your scripts is built into the core of Airflow using the powerful Jinja templating engine.
• Scalable: Airflow has a modular architecture and uses a message queue to orchestrate an arbitrary number of workers.

Requirements

Apache Airflow is tested with:

	Main version (dev)	Stable version (2.9.0)
Python	3.8, 3.9, 3.10, 3.11, 3.12	3.8, 3.9, 3.10, 3.11
Platform	AMD64/ARM64(*)	AMD64/ARM64(*)
Kubernetes	1.26, 1.27, 1.28, 1.29	1.25, 1.26, 1.27, 1.28, 1.29
PostgreSQL	12, 13, 14, 15, 16	12, 13, 14, 15, 16
MySQL	8.0, Innovation	8.0, Innovation
SQLite	3.15.0+	3.15.0+

* Experimental

Note: MySQL 5.x versions are unable to or have limitations with running multiple schedulers -- please see the Scheduler docs. MariaDB is not tested/recommended.

Note: SQLite is used in Airflow tests. Do not use it in production. We recommend using the latest stable version of SQLite for local development.

Note: Airflow currently can be run on POSIX-compliant Operating Systems. For development, it is regularly tested on fairly modern Linux Distros and recent versions of macOS. On Windows you can run it via WSL2 (Windows Subsystem for Linux 2) or via Linux Containers. The work to add Windows support is tracked via #10388, but it is not a high priority. You should only use Linux-based distros as "Production" execution environment as this is the only environment that is supported. The only distro that is used in our CI tests and that is used in the Community managed DockerHub image is Debian Bookworm. We also have support for legacy Debian Bullseye base image if you want to build a custom image but it is deprecated and option to do it will be removed in the Dockerfile that will accompany Airflow 2.9.0 so you are advised to switch to Debian Bookworm for your custom images.

Getting started

Visit the official Airflow website documentation (latest stable release) for help with installing Airflow, getting started, or walking through a more complete tutorial.

Note: If you're looking for documentation for the main branch (latest development branch): you can find it on s.apache.org/airflow-docs.

For more information on Airflow Improvement Proposals (AIPs), visit the Airflow Wiki.

Documentation for dependent projects like provider packages, Docker image, Helm Chart, you'll find it in the documentation index.

Installing from PyPI

We publish Apache Airflow as apache-airflow package in PyPI. Installing it however might be sometimes tricky because Airflow is a bit of both a library and application. Libraries usually keep their dependencies open, and applications usually pin them, but we should do neither and both simultaneously. We decided to keep our dependencies as open as possible (in pyproject.toml) so users can install different versions of libraries if needed. This means that pip install apache-airflow will not work from time to time or will produce unusable Airflow installation.

To have repeatable installation, however, we keep a set of "known-to-be-working" constraint files in the orphan constraints-main and constraints-2-0 branches. We keep those "known-to-be-working" constraints files separately per major/minor Python version. You can use them as constraint files when installing Airflow from PyPI. Note that you have to specify correct Airflow tag/version/branch and Python versions in the URL.

1. Installing just Airflow:

Note: Only pip installation is currently officially supported.

While it is possible to install Airflow with tools like Poetry or pip-tools, they do not share the same workflow as pip - especially when it comes to constraint vs. requirements management. Installing via Poetry or pip-tools is not currently supported.

There are known issues with bazel that might lead to circular dependencies when using it to install Airflow. Please switch to pip if you encounter such problems. Bazel community works on fixing the problem in this PR <https://github.com/bazelbuild/rules_python/pull/1166>_ so it might be that newer versions of bazel will handle it.

If you wish to install Airflow using those tools, you should use the constraint files and convert them to the appropriate format and workflow that your tool requires.

pip install 'apache-airflow==2.9.0' \
 --constraint "https://raw.githubusercontent.com/apache/airflow/constraints-2.9.0/constraints-3.8.txt"

2. Installing with extras (i.e., postgres, google)

pip install 'apache-airflow[postgres,google]==2.8.3' \
 --constraint "https://raw.githubusercontent.com/apache/airflow/constraints-2.9.0/constraints-3.8.txt"

For information on installing provider packages, check providers.

Official source code

Apache Airflow is an Apache Software Foundation (ASF) project, and our official source code releases:

• Follow the ASF Release Policy
• Can be downloaded from the ASF Distribution Directory
• Are cryptographically signed by the release manager
• Are officially voted on by the PMC members during the Release Approval Process

Following the ASF rules, the source packages released must be sufficient for a user to build and test the release provided they have access to the appropriate platform and tools.

Convenience packages

There are other ways of installing and using Airflow. Those are "convenience" methods - they are not "official releases" as stated by the ASF Release Policy, but they can be used by the users who do not want to build the software themselves.

Those are - in the order of most common ways people install Airflow:

• PyPI releases to install Airflow using standard pip tool
• Docker Images to install airflow via docker tool, use them in Kubernetes, Helm Charts, docker-compose, docker swarm, etc. You can read more about using, customising, and extending the images in the Latest docs, and learn details on the internals in the images document.
• Tags in GitHub to retrieve the git project sources that were used to generate official source packages via git

All those artifacts are not official releases, but they are prepared using officially released sources. Some of those artifacts are "development" or "pre-release" ones, and they are clearly marked as such following the ASF Policy.

User Interface

• DAGs: Overview of all DAGs in your environment.

  

• Grid: Grid representation of a DAG that spans across time.

  

• Graph: Visualization of a DAG's dependencies and their current status for a specific run.

  

• Task Duration: Total time spent on different tasks over time.

  

• Gantt: Duration and overlap of a DAG.

  

• Code: Quick way to view source code of a DAG.

  

Semantic versioning

As of Airflow 2.0.0, we support a strict SemVer approach for all packages released.

There are few specific rules that we agreed to that define details of versioning of the different packages:

• Airflow: SemVer rules apply to core airflow only (excludes any changes to providers). Changing limits for versions of Airflow dependencies is not a breaking change on its own.
• Airflow Providers: SemVer rules apply to changes in the particular provider's code only. SemVer MAJOR and MINOR versions for the packages are independent of the Airflow version. For example, google 4.1.0 and amazon 3.0.3 providers can happily be installed with Airflow 2.1.2. If there are limits of cross-dependencies between providers and Airflow packages, they are present in providers as install_requires limitations. We aim to keep backwards compatibility of providers with all previously released Airflow 2 versions but there will sometimes be breaking changes that might make some, or all providers, have minimum Airflow version specified.
• Airflow Helm Chart: SemVer rules apply to changes in the chart only. SemVer MAJOR and MINOR versions for the chart are independent of the Airflow version. We aim to keep backwards compatibility of the Helm Chart with all released Airflow 2 versions, but some new features might only work starting from specific Airflow releases. We might however limit the Helm Chart to depend on minimal Airflow version.
• Airflow API clients: Their versioning is independent from Airflow versions. They follow their own SemVer rules for breaking changes and new features - which for example allows to change the way we generate the clients.

Version Life Cycle

Apache Airflow version life cycle:

Version	Current Patch/Minor	State	First Release	Limited Support	EOL/Terminated
2	2.9.0	Supported	Dec 17, 2020	TBD	TBD
1.10	1.10.15	EOL	Aug 27, 2018	Dec 17, 2020	June 17, 2021
1.9	1.9.0	EOL	Jan 03, 2018	Aug 27, 2018	Aug 27, 2018
1.8	1.8.2	EOL	Mar 19, 2017	Jan 03, 2018	Jan 03, 2018
1.7	1.7.1.2	EOL	Mar 28, 2016	Mar 19, 2017	Mar 19, 2017

Limited support versions will be supported with security and critical bug fix only. EOL versions will not get any fixes nor support. We always recommend that all users run the latest available minor release for whatever major version is in use. We highly recommend upgrading to the latest Airflow major release at the earliest convenient time and before the EOL date.

Support for Python and Kubernetes versions

As of Airflow 2.0, we agreed to certain rules we follow for Python and Kubernetes support. They are based on the official release schedule of Python and Kubernetes, nicely summarized in the Python Developer's Guide and Kubernetes version skew policy.

1. We drop support for Python and Kubernetes versions when they reach EOL. Except for Kubernetes, a version stays supported by Airflow if two major cloud providers still provide support for it. We drop support for those EOL versions in main right after EOL date, and it is effectively removed when we release the first new MINOR (Or MAJOR if there is no new MINOR version) of Airflow. For example, for Python 3.8 it means that we will drop support in main right after 27.06.2023, and the first MAJOR or MINOR version of Airflow released after will not have it.

2. We support a new version of Python/Kubernetes in main after they are officially released, as soon as we make them work in our CI pipeline (which might not be immediate due to dependencies catching up with new versions of Python mostly) we release new images/support in Airflow based on the working CI setup.

3. This policy is best-effort which means there may be situations where we might terminate support earlier if circumstances require it.

Base OS support for reference Airflow images

The Airflow Community provides conveniently packaged container images that are published whenever we publish an Apache Airflow release. Those images contain:

• Base OS with necessary packages to install Airflow (stable Debian OS)
• Base Python installation in versions supported at the time of release for the MINOR version of Airflow released (so there could be different versions for 2.3 and 2.2 line for example)
• Libraries required to connect to supported Databases (again the set of databases supported depends on the MINOR version of Airflow)
• Predefined set of popular providers (for details see the Dockerfile).
• Possibility of building your own, custom image where the user can choose their own set of providers and libraries (see Building the image)
• In the future Airflow might also support a "slim" version without providers nor database clients installed

The version of the base OS image is the stable version of Debian. Airflow supports using all currently active stable versions - as soon as all Airflow dependencies support building, and we set up the CI pipeline for building and testing the OS version. Approximately 6 months before the end-of-regular support of a previous stable version of the OS, Airflow switches the images released to use the latest supported version of the OS.

For example since Debian Buster end-of-life was August 2022, Airflow switched the images in main branch to use Debian Bullseye in February/March 2022. The version was used in the next MINOR release after the switch happened. In case of the Bullseye switch - 2.3.0 version used Debian Bullseye. The images released in the previous MINOR version continue to use the version that all other releases for the MINOR version used. Similar switch from Debian Bullseye to Debian Bookworm has been implemented before 2.8.0 release in October 2023 and Debian Bookworm will be the only option supported as of Airflow 2.9.0.

Users will continue to be able to build their images using stable Debian releases until the end of regular support and building and verifying of the images happens in our CI but no unit tests were executed using this image in the main branch.

Approach to dependencies of Airflow

Airflow has a lot of dependencies - direct and transitive, also Airflow is both - library and application, therefore our policies to dependencies has to include both - stability of installation of application, but also ability to install newer version of dependencies for those users who develop DAGs. We developed the approach where constraints are used to make sure airflow can be installed in a repeatable way, while we do not limit our users to upgrade most of the dependencies. As a result we decided not to upper-bound version of Airflow dependencies by default, unless we have good reasons to believe upper-bounding them is needed because of importance of the dependency as well as risk it involves to upgrade specific dependency. We also upper-bound the dependencies that we know cause problems.

The constraint mechanism of ours takes care about finding and upgrading all the non-upper bound dependencies automatically (providing that all the tests pass). Our main build failures will indicate in case there are versions of dependencies that break our tests - indicating that we should either upper-bind them or that we should fix our code/tests to account for the upstream changes from those dependencies.

Whenever we upper-bound such a dependency, we should always comment why we are doing it - i.e. we should have a good reason why dependency is upper-bound. And we should also mention what is the condition to remove the binding.

Approach for dependencies for Airflow Core

Those dependencies are maintained in pyproject.toml.

There are few dependencies that we decided are important enough to upper-bound them by default, as they are known to follow predictable versioning scheme, and we know that new versions of those are very likely to bring breaking changes. We commit to regularly review and attempt to upgrade to the newer versions of the dependencies as they are released, but this is manual process.

The important dependencies are:

• SQLAlchemy: upper-bound to specific MINOR version (SQLAlchemy is known to remove deprecations and introduce breaking changes especially that support for different Databases varies and changes at various speed)
• Alembic: it is important to handle our migrations in predictable and performant way. It is developed together with SQLAlchemy. Our experience with Alembic is that it very stable in MINOR version
• Flask: We are using Flask as the back-bone of our web UI and API. We know major version of Flask are very likely to introduce breaking changes across those so limiting it to MAJOR version makes sense
• werkzeug: the library is known to cause problems in new versions. It is tightly coupled with Flask libraries, and we should update them together
• celery: Celery is a crucial component of Airflow as it used for CeleryExecutor (and similar). Celery follows SemVer, so we should upper-bound it to the next MAJOR version. Also, when we bump the upper version of the library, we should make sure Celery Provider minimum Airflow version is updated.
• kubernetes: Kubernetes is a crucial component of Airflow as it is used for the KubernetesExecutor (and similar). Kubernetes Python library follows SemVer, so we should upper-bound it to the next MAJOR version. Also, when we bump the upper version of the library, we should make sure Kubernetes Provider minimum Airflow version is updated.

Approach for dependencies in Airflow Providers and extras

The main part of the Airflow is the Airflow Core, but the power of Airflow also comes from a number of providers that extend the core functionality and are released separately, even if we keep them (for now) in the same monorepo for convenience. You can read more about the providers in the Providers documentation. We also have set of policies implemented for maintaining and releasing community-managed providers as well as the approach for community vs. 3rd party providers in the providers document.

Those extras and providers dependencies are maintained in provider.yaml of each provider.

By default, we should not upper-bound dependencies for providers, however each provider's maintainer might decide to add additional limits (and justify them with comment).

Contributing

Want to help build Apache Airflow? Check out our contributing documentation.

Official Docker (container) images for Apache Airflow are described in images.

Voting Policy

• Commits need a +1 vote from a committer who is not the author
• When we do AIP voting, both PMC member's and committer's +1s are considered a binding vote.

Who uses Apache Airflow?

We know about around 500 organizations that are using Apache Airflow (but there are likely many more) in the wild.

If you use Airflow - feel free to make a PR to add your organisation to the list.

Who maintains Apache Airflow?

Airflow is the work of the community, but the core committers/maintainers are responsible for reviewing and merging PRs as well as steering conversations around new feature requests. If you would like to become a maintainer, please review the Apache Airflow committer requirements.

What goes into the next release?

Often you will see an issue that is assigned to specific milestone with Airflow version, or a PR that gets merged to the main branch and you might wonder which release the merged PR(s) will be released in or which release the fixed issues will be in. The answer to this is as usual - it depends on various scenarios. The answer is different for PRs and Issues.

To add a bit of context, we are following the Semver versioning scheme as described in Airflow release process. More details are explained in detail in this README under the Semantic versioning chapter, but in short, we have MAJOR.MINOR.PATCH versions of Airflow.

• MAJOR version is incremented in case of breaking changes
• MINOR version is incremented when there are new features added
• PATCH version is incremented when there are only bug-fixes and doc-only changes

Generally we release MINOR versions of Airflow from a branch that is named after the MINOR version. For example 2.7.* releases are released from v2-7-stable branch, 2.8.* releases are released from v2-8-stable branch, etc.

Most of the time in our release cycle, when the branch for next MINOR branch is not yet created, all PRs merged to main (unless they get reverted), will find their way to the next MINOR release. For example if the last release is 2.7.3 and v2-8-stable branch is not created yet, the next MINOR release is 2.8.0 and all PRs merged to main will be released in 2.8.0. There is a brief period of time when we cut a new MINOR release branch and prepare the alpha, beta, RC candidates for the 2.NEXT_MINOR.0 version where PRs merged to main will only be released in the following MINOR release.

However, some PRs (bug-fixes and doc-only changes) when merged, can be cherry-picked to current MINOR branch and released in the next PATCHLEVEL release - for example when the last released version from v2-7-stable branch is 2.7.2. Some of the PRs from main can be marked as 2.7.3 milestone by committers, the release manager will try to cherry-pick them into the release branch. If successful, they will be released in 2.7.3. The final decision about cherry-picking is made by the release manager.

Marking issues with a milestone is a bit different. Maintainers do not mark issues with a milestone usually, normally they are only marked in PRs. If PR linked to the issue (and "fixing it") gets merged and released in a specific version following the process described above, the issue will be automatically closed, no milestone will be set for the issue, you need to check the PR that fixed the issue to see which version it was released in.

However, sometimes maintainers mark issues with specific milestone, which means that the issue is important to become a candidate to take a look when the release is being prepared. Since this is an Open-Source project, where basically all contributors volunteer their time, there is no guarantee that specific issue will be fixed in specific version. We do not want to hold the release because some issue is not fixed, so in such case release manager will reassign such unfixed issues to the next milestone in case they are not fixed in time for the current release. Therefore, the milestone for issue is more of an intent that it should be looked at, than promise it will be fixed in the version.

More context and FAQ about the patchlevel release can be found in the What goes into the next release document in the dev folder of the repository.

Can I use the Apache Airflow logo in my presentation?

Yes! Be sure to abide by the Apache Foundation trademark policies and the Apache Airflow Brandbook. The most up-to-date logos are found in this repo and on the Apache Software Foundation website.

Airflow merchandise

If you would love to have Apache Airflow stickers, t-shirt, etc. then check out Redbubble Shop.

Links

• Documentation
• Chat
• Community Information

Sponsors

The CI infrastructure for Apache Airflow has been sponsored by:

airflow's Issues

Run a DAG from command line

Is there a way to run a DAG from command line?

I see two options:

1. Use backfill and give same start and end date.
2. Use run - possible only if task_id is optional. Currently, run seems to support only running a specific task within a specific DAG. Is it possible to make task_id optional, so that if it's not provide, the run command will run the full DAG?

Is there any other way of doing this that I may have missed in the documentation?

New Dags are only imported when the webserver restarts!

How does one add new dags to the system without restarting the webserver? Running "python my_dag.py" is not importing the dag in to the db either. I am running on EC2 with a Postgresql database. I am able to see the data gets imported only on restart of the web app.

Here is the code:

This code will become the EP Data Pipeline.

The flow may eventually look like : 
* 1. SQS Message Written Detect
* 2.a. Email Send Flow Started 
* 2.b. Database Row Written Detect
* 3. SQS Queue Empty Detect
* 4. Email Send Flow Complete 
"""
from airflow import DAG
from airflow.operators import EmailOperator
from datetime import datetime


print ' got here 1'
default_args = {
    'owner': 'airflow',
    'depends_on_past': False,
    'start_date': datetime(2015, 1, 1),
    'email': ['[email protected]'],
    'email_on_failure': True,
    'email_on_retry': True,
}

print ' got here 1'
dag = DAG('ep_demo_1', default_args=default_args)

print ' got here 1'
t1 = EmailOperator(
    task_id='email_pipeline_start',
    to='[email protected]',
    subject='EP Demo Pipeline Started (TAD)',
    html_content='',
    dag=dag)

After restarting the webapp, I still don't see any entries in the dag table in the sqlite db for my dags. I see the dags in the UI -- there seems to be a difference between loading and importing a dag.

2015-06-18 17:49:21,175 - root - INFO - Loaded DAG <DAG: ep_demo_1>
2015-06-18 17:49:21,177 - root - INFO - Importing /usr/local/lib/python2.7/site-packages/airflow/example_dags/example_bash_operator.py
2015-06-18 17:49:21,178 - root - INFO - Loaded DAG <DAG: example_bash_operator>
2015-06-18 17:49:21,179 - root - INFO - Importing /usr/local/lib/python2.7/site-packages/airflow/example_dags/example_python_operator.py
2015-06-18 17:49:21,180 - root - INFO - Loaded DAG <DAG: example_python_operator>
2015-06-18 17:49:21,181 - root - INFO - Importing /usr/local/lib/python2.7/site-packages/airflow/example_dags/tutorial.py
2015-06-18 17:49:21,182 - root - INFO - Loaded DAG <DAG: tutorial>

Understanding interactions between DAG scheduling and sensors

Let's say I have a DAG scheduled with a one-hour schedule interval, and the first step in that DAG is a sensor that pokes once a minute.

What happens if the sensor never triggers during the hour after the first DAG "submission"? What state are tasks in the first DAG left in? What if the sensor's triggering event does occur at some point after the second DAG is scheduled? Does that "release" the remainder of both DAGs?

Or maybe the example I propose just doesn't make sense as a way to set things up. If so, could you explain why?

I'm sure this also somehow interacts with depends_on_past, but I feel like I'd be reaching too far to try and guess how. (This might be a candidate for documentation expansion.)

Airflow Worker not Working When I run Backfill commands

I am using the CeleryExecutor with a Postgres results backend and queue. It looks like this fails when I run "airflow backfill" commands but works when I use the UI.

Sid-As-MBP-15:airflow siddharth$ airflow worker

 -------------- [email protected] v3.1.18 (Cipater)
---- **** ----- 
--- * ***  * -- Darwin-14.3.0-x86_64-i386-64bit
-- * - **** --- 
- ** ---------- [config]
- ** ---------- .> app:         airflow.executors.celery_executor:0x10c04e690
- ** ---------- .> transport:   sqla+postgresql://siddharth@localhost:5432/airflow_celery
- ** ---------- .> results:     db+postgresql://siddharth@localhost:5432/airflow_celery
- *** --- * --- .> concurrency: 16 (prefork)
-- ******* ---- 
--- ***** ----- [queues]
 -------------- .> default          exchange=celery(direct) key=celery


[2015-06-22 15:08:15,337: WARNING/MainProcess] [email protected] ready.
Starting flask
 * Running on http://0.0.0.0:8793/ (Press CTRL+C to quit)
Traceback (most recent call last):
  File "/usr/local/bin/airflow", line 10, in <module>
    args.func(args)
  File "/usr/local/lib/python2.7/site-packages/airflow/bin/cli.py", line 97, in run
    DagPickle).filter(DagPickle.id == args.pickle).first()
  File "/usr/local/lib/python2.7/site-packages/sqlalchemy/orm/query.py", line 2444, in first
    ret = list(self[0:1])
  File "/usr/local/lib/python2.7/site-packages/sqlalchemy/orm/query.py", line 2280, in __getitem__
    return list(res)
  File "/usr/local/lib/python2.7/site-packages/sqlalchemy/orm/loading.py", line 84, in instances
    util.raise_from_cause(err)
  File "/usr/local/lib/python2.7/site-packages/sqlalchemy/util/compat.py", line 199, in raise_from_cause
    reraise(type(exception), exception, tb=exc_tb)
  File "/usr/local/lib/python2.7/site-packages/sqlalchemy/orm/loading.py", line 69, in instances
    rows = [proc(row) for row in fetch]
  File "/usr/local/lib/python2.7/site-packages/sqlalchemy/orm/loading.py", line 426, in _instance
    loaded_instance, populate_existing, populators)
  File "/usr/local/lib/python2.7/site-packages/sqlalchemy/orm/loading.py", line 484, in _populate_full
    dict_[key] = getter(row)
  File "/usr/local/lib/python2.7/site-packages/sqlalchemy/sql/sqltypes.py", line 1261, in process
    return loads(value)
  File "/usr/local/lib/python2.7/site-packages/dill/dill.py", line 209, in loads
    return load(file)
  File "/usr/local/lib/python2.7/site-packages/dill/dill.py", line 199, in load
    obj = pik.load()
  File "/usr/local/Cellar/python/2.7.9/Frameworks/Python.framework/Versions/2.7/lib/python2.7/pickle.py", line 858, in load
    dispatch[key](self)
  File "/usr/local/Cellar/python/2.7.9/Frameworks/Python.framework/Versions/2.7/lib/python2.7/pickle.py", line 1090, in load_global
    klass = self.find_class(module, name)
  File "/usr/local/lib/python2.7/site-packages/dill/dill.py", line 278, in find_class
    return StockUnpickler.find_class(self, module, name)
  File "/usr/local/Cellar/python/2.7.9/Frameworks/Python.framework/Versions/2.7/lib/python2.7/pickle.py", line 1124, in find_class
    __import__(module)
ImportError: No module named python_operator
[2015-06-22 15:11:08,787: ERROR/Worker-14] 1
[2015-06-22 15:11:08,835: ERROR/MainProcess] Task airflow.executors.celery_executor.execute_command[4a012cf5-83f6-4d54-af34-b2ce74b517db] raised unexpected: Exception('Celery command failed',)
Traceback (most recent call last):
  File "/usr/local/lib/python2.7/site-packages/celery/app/trace.py", line 240, in trace_task
    R = retval = fun(*args, **kwargs)
  File "/usr/local/lib/python2.7/site-packages/celery/app/trace.py", line 438, in __protected_call__
    return self.run(*args, **kwargs)
  File "/usr/local/lib/python2.7/site-packages/airflow/executors/celery_executor.py", line 41, in execute_command
    raise Exception('Celery command failed')
Exception: Celery command failed

My airflow.cfg file has the following settings for celery

# The Celery broker URL. Celery supports RabbitMQ, Redis and experimentaly
# a sqlalchemy database. Refer to the Celery documentation for more
# information.
#broker_url = sqla+mysql://airflow:airflow@localhost:3306/airflow
broker_url = sqla+postgresql://siddharth@localhost:5432/airflow_celery

# Another key Celery setting
#celery_result_backend = db+mysql://airflow:airflow@localhost:3306/airflow
celery_result_backend = db+postgresql://siddharth@localhost:5432/airflow_celery

I am getting this error with the example_python_operator and example_bash_operator flows. I haven't tried others.

airflow flower is not much help for experimental brokers like Postgres (over SQLAlchemy) since many commands are unsupported as mentioned in the Limitations section of this

Sid-As-MacBook-Pro-15:airflow siddharth$ airflow flower
[I 150622 15:05:54 command:114] Visit me at http://localhost:5555
[I 150622 15:05:55 command:116] Broker: sqla+postgresql://siddharth@localhost:5432/airflow_celery
[I 150622 15:05:55 command:119] Registered tasks:
    ['celery.backend_cleanup',
     'celery.chain',
     'celery.chord',
     'celery.chord_unlock',
     'celery.chunks',
     'celery.group',
     'celery.map',
     'celery.starmap']
[I 150622 15:05:55 mixins:225] Connected to sqla+postgresql://siddharth@localhost:5432/airflow_celery
[W 150622 15:05:58 control:43] 'stats' inspect method failed
[W 150622 15:05:58 control:43] 'active_queues' inspect method failed
[W 150622 15:05:58 control:43] 'registered' inspect method failed
[W 150622 15:05:58 control:43] 'scheduled' inspect method failed
[W 150622 15:05:58 control:43] 'active' inspect method failed
[W 150622 15:05:58 control:43] 'reserved' inspect method failed
[W 150622 15:05:58 control:43] 'revoked' inspect method failed
[W 150622 15:05:58 control:43] 'conf' inspect method failed

If I look into my database (airflow_celery), I find 2 tables : kombu_message and kombu_queue.

In kombu_queue, I'm expecting a single queue called default, but I see several

airflow_celery=# select * from kombu_queue ;
 id |                           name                           
----+----------------------------------------------------------
  1 | default
  2 | [email protected]
  3 | celeryev.ad9d88a9-05b7-4d01-ad2c-8f95662787a0
  4 | 37bb18ec-a47f-349b-859f-3b454d2602a3.reply.celery.pidbox
  5 | b15e35ef-8873-31d6-8841-af833b9cec3d.reply.celery.pidbox
  6 | 4572302e-65a9-3ec1-850e-aff27cd59a9b.reply.celery.pidbox
  7 | 33d4dcc5-f997-3b7c-9f41-66d9a2c51159.reply.celery.pidbox

The query stuck for hive2

I used a simple query for a hive2 connection. It have been sucked. I cannot get the result ;(

2015-06-19 10:20:56,647 - tornado.access - INFO - 200 GET /admin/queryview/ (10.196.11.238) 19.21ms
2015-06-19 10:21:02,962 - root - INFO - Using connection to: localhost
2015-06-19 10:21:02,965 - root - INFO - Using connection to: localhost

Scheduling frequency

Is there a way to schedule jobs/dags on specific time of the day? Also, on a related note, can DAGs be scheduled on some other frequency - such as hourly, weekly etc?

Apologies if I missed it in the documentation.

Can't add external executors -> Plugin manager

Would be nice to be able to add other executors "out of airflow codebase".

List of executors is hard coded in airflow/executors/init.py

A kinda plugin mechanism could allow to add other executors.

airflow worker does not work

$ airflow worker
Traceback (most recent call last):
File "/usr/local/bin/airflow", line 10, in
args.func(args)
File "/usr/local/lib/python2.7/site-packages/airflow/bin/cli.py", line 299, in worker
worker.run(**options)
File "/usr/local/lib/python2.7/site-packages/celery/bin/worker.py", line 212, in run
state_db=self.node_format(state_db, hostname), **kwargs
File "/usr/local/lib/python2.7/site-packages/celery/worker/init.py", line 100, in init
self.setup_instance(**self.prepare_args(**kwargs))
File "/usr/local/lib/python2.7/site-packages/celery/worker/init.py", line 124, in setup_instance
self._conninfo = self.app.connection()
File "/usr/local/lib/python2.7/site-packages/celery/app/base.py", line 385, in connection
'BROKER_CONNECTION_TIMEOUT', connect_timeout
File "/usr/local/lib/python2.7/site-packages/kombu/connection.py", line 165, in init
if not get_transport_cls(transport).can_parse_url:
File "/usr/local/lib/python2.7/site-packages/kombu/transport/init.py", line 109, in get_transport_cls
_transport_cache[transport] = resolve_transport(transport)
File "/usr/local/lib/python2.7/site-packages/kombu/transport/init.py", line 89, in resolve_transport
raise KeyError('No such transport: {0}'.format(transport))
KeyError: 'No such transport: '

Add `only_run_latest` to BaseOperator

Dependencies on an only_run_latest task are satisfied when that task has succeeded at any point in the future of the task depending on it.

When backfilling, only the end_date of only_run_latest tasks would be triggered, and dependencies for this task through time would be satisfied upon completion of this latest task instance.

When it comes to an only_run_latest task, the scheduler should skip runs to max(execution_date) of all of its dependencies.

Gantt Chart Bug

Here is my flow.

I'm noticing something wrong in the gantt chart.

The order of the first 2 stages in my gantt chart does not reflect my stage dependencies.
So I run "submit_spark_job" followed by "wait_for_first_sqs_message". Why is it then that the gantt chart shows them out of order?

My code is below:

from airflow import DAG
from airflow.operators import PythonOperator, PostgresOperator, BashOperator
from datetime import date, datetime, time, timedelta
from pprint import pprint

import base64
import boto.sqs
import boto.sns
import json
import logging
import psycopg2
import sys
import time

# CONSTANTS
SQS_QUEUE_NAME='*******'
REGION = 'us-west-2'
DB_CHECK_QUERY='SELECT count(*) from message;'
SNS_TOPIC_ARN_FOR_JOB_STATUS='*******'
EP_DB_CONNECT_STRING = "*******"


default_args = {
    'owner': 'airflow',
    'depends_on_past': False,
    'start_date': datetime(2015, 6, 22),
    'schedule_interval': timedelta(days=1),
    'email': ['[email protected]'],
    'email_on_failure': True,
    'email_on_retry': True,
    'retries': 3,
    'retry_interval': timedelta(seconds=5)
}
dag = DAG('ep_demo', default_args=default_args)

def _wait_for_first_sqs_message_internal(ds, **kwargs):
    conn = boto.sqs.connect_to_region(REGION)
    q = conn.get_queue(SQS_QUEUE_NAME)
    rs = q.get_messages()
    while rs is None or len(rs) <= 0:
        logging.info("----- WAITING FOR A MESSAGE on SQS Queue : {}".format(SQS_QUEUE_NAME))
        time.sleep(10)
        rs = q.get_messages()
    logging.info("----- FOUND A MESSAGE on SQS Queue : {}".format(SQS_QUEUE_NAME))
    return True

def _wait_for_empty_queue_internal(ds, **kwargs):
    conn = boto.sqs.connect_to_region(REGION)
    q = conn.get_queue(SQS_QUEUE_NAME)
    q_size = q.count()
    while q_size > 0:
        logging.info("----- WAITING FOR Queue {} to Drain".format(SQS_QUEUE_NAME))
        q = conn.get_queue(SQS_QUEUE_NAME)
        q_size = q.count()
        time.sleep(10)
    logging.info("----- Queue {} Empty".format(SQS_QUEUE_NAME))
    return True

# TODO : pointing to local db, parameterize this
def _wait_for_new_data_in_db_internal(ds, **kwargs):
    db_conn = psycopg2.connect(EP_DB_CONNECT_STRING)
    logging.info("----- Successfully Connected to database {}".format(EP_DB_CONNECT_STRING))
    cursor = db_conn.cursor()

    # Establish a base line
    cursor.execute(DB_CHECK_QUERY)
    result = cursor.fetchone()
    initial_row_count = int(result[0])
    found_new_data = None

    # Detect a change in the record count
    while not found_new_data:
        logging.info("----- WAITING FOR A NEW DATA in DB : base count = {}".format(initial_row_count))
        time.sleep(10)
        cursor.execute(DB_CHECK_QUERY)
        result = cursor.fetchone()
        new_row_count = int(result[0])
        if(new_row_count > initial_row_count):
            found_new_data = True
    logging.info("----- FOUND A NEW DATA in DB : (base count ==> new count) = ({} ==> {})".format(new_row_count, initial_row_count))
    return True


def _send_sns_notification_internal(ds, **kwargs):
    conn = boto.sns.connect_to_region(REGION)
    conn.publish(message='Airflow EP Data pipeline flow completed successfully', subject='Airflow Flow Complete', target_arn=SNS_TOPIC_ARN_FOR_JOB_STATUS,
                )
    logging.info("----- Sent SNS about FLOW Completion")
    return True

# Operator 0 : Start a spark job
#submit_spark_job = BashOperator(
#    task_id='submit_spark_job',
#    bash_command="ssh -i ~/.ssh/id_rsa_spark root@*******'. /root/.bash_profile; cd $SPARK_HOME/bin; ~/spark/bin/spark-submit --py-files risk.py,volume_dict.py,domain_suffix.py,sender_models.py,avro_tools.py aggregate.py --avro --input uploads/1/db80b4a2-fda4-11e4-9f6c-062f18616d35/ --s3 --domains --bucket agari-collector-ingest-avro'",
#    dag=dag)

submit_spark_job = BashOperator(
    task_id='submit_spark_job',
    bash_command="ssh -i ~/.ssh/id_rsa_spark root@*******'. /root/.bash_profile; ~/run_ingest.sh'",
    dag=dag)

# Operator 1 : Check a SQS queue for messages - currently a dummy!
wait_for_first_sqs_message = PythonOperator(
    task_id='wait_for_first_sqs_message',
    provide_context=True,
    python_callable=_wait_for_first_sqs_message_internal,
    dag=dag)
wait_for_first_sqs_message.set_upstream(submit_spark_job)

# Operator 2.a : Check for new data in DB
wait_for_new_data_in_db = PythonOperator(
    task_id='wait_for_new_data_in_db',
    provide_context=True,
    python_callable=_wait_for_new_data_in_db_internal,
    dag=dag)
wait_for_new_data_in_db.set_upstream(wait_for_first_sqs_message)


# Operator 3 : Check for SQS Queue to be drained
wait_for_empty_queue = PythonOperator(
    task_id='wait_for_empty_queue',
    provide_context=True,
    python_callable=_wait_for_empty_queue_internal,
    dag=dag)
wait_for_empty_queue.set_upstream(wait_for_new_data_in_db)

# Operator 4b : Send SNS when flow completes successfully
send_sns_notification = PythonOperator(
    task_id='send_sns_notification',
    provide_context=True,
    python_callable=_send_sns_notification_internal,
    dag=dag)
send_sns_notification.set_upstream(wait_for_empty_queue)

Documentation guidelines

We'd like to use comments within our dags to generate job specific documentation.

This will enable us to version control, maintain and easily produce maintainable and understandable DAGs without the overheads of writing separate documents.

This seems fairly straight forward using docstring conventions. However, I'd like to provide useful comments that explain the logic within each task within a DAG. Is there a way to use sphinx (or something similar) to generate documentation? It would be brilliant if we could produce html/pdf/any format documentation that detailed each task and its dependencies/inputs etc..

Why do we need a CeleryExecutor in order to use the UI features?

It seems the start up costs of this are a bit high. I am trying out Luigi and Airflow, and need to set up a central scheduler for both> The UI features for Airflow are more attractive than those of Luigi, but the UI features in Airflow are only usable (actionable) if Celery is installed. I am running on EC2 and don't want to install too much to just take this for a spin. My recommendation is to make the UI features work for the LocalExecutor. Otherwise, Luigi seems attractive -- the startup cost is similar but it is more battletested

-s

New feature: HTTP interface?

So I went through the codebase and docs of 'airflow' today and I think it's a great fit for one of my projects. I'm the maintainer of "remap", which is a 100% python implementation of MapReduce only intended to run on a dozen of nodes for now. Jobs are kicked off through a REST interface.

Here's where a potential contribution comes in.

I didn't find anything already done with http interfaces, so my idea is to write an HTTPHook, an HTTP operator and a sensor for this. The operator/hook calls a URL resource with an indicated method and potentially some post data. The sensor later on calls another URL to check on progress. This would allow work to be executed asynchronously until some later checkpoint where the sensor needs to check whether something is available or not.

The HTTP library I intend to use is "requests", which should come installed with "pip": http://docs.python-requests.org/en/latest/

Operators never seem to return values and probably by design, which means that a worker process waits around for the job to complete, which means that an operator executes the action synchronously.

So, couple of questions:

1. Is this a welcome contribution? Is someone already working on an HTTP hook?
2. Can operators return data that is checked for later with a sensor to support inherently asynchronous systems like this one? Or are operators required to be synchronous in the sense that it always waits for the result of an operation, so must keep checking inside the operator at intervals until the action completes?
3. Is the "requests" library an acceptable library (considering license, availability, installability, etc)?

As extra thought on 2, it's possible that external systems contain key/values that are useful to use in workflows. Is there a recommended mechanism of loading small pieces of data into a DAG workflow so that it's available from a context for example when another task is executed?

How to execute arbitrary SQL in postgres?

I'd like to execute a series of SQL statements on Redshift.

I'd planned to use the SqlSensor to do this but failed to read the part of the documentation about executing a statement until a criteria is met. I just need to execute regardless of the result set.

Is there a way of doing this without using the PythonOperator and opening my own postgres connection? It looks as if I want a MySqlOperator for postgres.

Disabling backfill functionality?

If I have a process that needs to run daily regardless of input parameters, for example a process that bulk loads a table and doesn't depend on delta data, is there a way of preventing backfills running a dag multiple times?

If I have a dag with for arguments sake, one task, can I make a backfill with a weeks' date range only run the task once rather than 7 times?

Right way to shutdown the scheduler and webserver

I'm integrating Airflow with monit.

Webserver
When you type airflow webserver -d -p 8080, 2 webserver processes are launched, one is a child of the other.

deploy   19005     1  0 00:37 ?        00:00:01 /usr/bin/python /usr/local/bin/airflow webserver -d -p 808 
deploy   19010 19005  1 00:37 ?        00:00:17 /usr/bin/python /usr/local/bin/airflow webserver -d -p 808 

In order to correctly kill a process tree with 2 levels (parent and child), I use the "stop program" command in the monit conf.d file below :

check process airflow-webserver with pidfile /home/deploy/airflow/pids/airflow-webserver.pid
  group airflow-webserver
  start program "/bin/sh -c '( TMP=/data/tmp PATH=/bin:/sbin:/usr/bin:/usr/sbin:/usr/local/bin airflow webserver -d -p 8080 2>&1 & echo $! > /home/deploy/airflow/pids/airflow-webserver.pid ) | logger -p user.info'"
    as uid deploy and gid deploy
  stop program "/bin/sh -c 'PATH=/bin:/sbin:/usr/bin:/usr/sbin pkill -TERM -P `cat /home/deploy/airflow/pids/airflow-webserver.pid` && rm -f /home/deploy/airflow/pids/airflow-webserver.pid'"
    as uid deploy and gid deploy

Ideally, airflow should clean up its own child processes when it receives a SIGTERM. But, since it doesn't, I need to call pkill with the -P option.

Scheduler
The Scheduler also launches a bunch of processes. However, for some reason, using the same approach, I always find the parent surviving.

check process airflow-scheduler with pidfile /home/deploy/airflow/pids/airflow-scheduler.pid
  group airflow-scheduler
  start program "/bin/sh -c '( TMP=/data/tmp PATH=/bin:/sbin:/usr/bin:/usr/sbin:/usr/local/bin airflow scheduler 2>&1 & echo $! > /home/deploy/airflow/pids/airflow-scheduler.pid ) | logger -p user.info'"
    as uid deploy and gid deploy
  stop program "/bin/sh -c 'PATH=/bin:/sbin:/usr/bin:/usr/sbin pkill -KILL -P `cat /home/deploy/airflow/pids/airflow-scheduler.pid` && rm -f /home/deploy/airflow/pids/airflow-scheduler.pid'"
    as uid deploy and gid deploy

Can you provide guidance on a clean way to shutdown the scheduler and webserver?

-s

'ascii' codec can't decode byte

Running through the Quickstart page right about the line that says:

pip install airflow

I get the following error:

Command python setup.py egg_info failed with error code 1 in /tmp/pip_build_hugo/pandas
Traceback (most recent call last):
  File "/usr/bin/pip", line 9, in <module>
    load_entry_point('pip==1.5.4', 'console_scripts', 'pip')()
  File "/usr/lib/python2.7/dist-packages/pip/__init__.py", line 235, in main
    return command.main(cmd_args)
  File "/usr/lib/python2.7/dist-packages/pip/basecommand.py", line 161, in main
    text = '\n'.join(complete_log)
UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 72: ordinal not in range(128)

It happens both on ubuntu 14.04 and Centos 7 at least...

Incorrect license in setup.cfg

The setup.cfg mentions an MIT license but the LICENSE.txt states the project uses Apache 2.0.

How Do I Execute Conditional Branching?

Let's say that I have 2 paths with a branching point.

a --> b1 --> c --> e --> f --> g
a --> b2 --> d

So, starting at "a", based on some condition, I either follow b1 --> c --> e --> f --> g or b2-->d. In my specific case, if there is data generated from stage "a", I expect to go through full processing of the data c --> e --> f --> g. If I don't have data, then I expect to execute "d", which simply emails that there is no new data to process today!

If the entire flow were executed in a single subprocess within a single python interpreter, I could set a global boolean (tsk tsk, I know) that all operators check in order to decide whether to "do something" or not. Essentially, I would be placing the conditional logic in all operators downstream of "b". I am using the LocalExecutor right now, but don't expect that this assumption will hold.

So, I would need to store the global in a database and have all downstream operators check that boolean. Basically, I would rather just have a "branch operator" instead, so that I don't need to do this! In my flow, "b' is the branch operator, with "b1" and "b2" as branches.

In the absence of a conditional operator, I am considering the following:
If there is no data, "c" will raise an error so that it is marked red and the downstream operators would not be executed. On the other branch, "d" would pass, sending an email that there is no new data.

If there is new data, then the opposite could happen.

My approach above will achieve what I need by showing one branch as having one "red" or "failed" task (i.e. at "c" or "d") and everything downstream as not being executed. That seems to meet my need

Memory usage

I think is quite linked with one of my previous questions, but just to be sure. Will you look at the memory usage of each tasks and output an awesome graph or things like that?

That could be very cool, I did not see any scheduler with a good ui doing that.

how to use variables ?

webserver presents a possibility to define variables but there is no documentation on how to use those variables (and I don't see any usage in code).

Is it a way to defined "constants" to use in templates/arguments? How can we use them?

Add an FTP hook

Do you have plans to support FTP hooks?

I've noticed there's an FTP connection type in the front-end, but I've not found a component to utilise the connection. Is this a case of me extending BaseHook to wrap some functionality to pull and push files via FTP or is there something existing that I've missed?

Airflow Tutorial page inconsistencies

I noticed some inconsistencies in your Tutorial page where you walk us through the DAG. For instance, in the code block, you have:

default_args = {
    'owner': 'airflow',
    'depends_on_past': False,
    'start_date': datetime(2015, 6, 1),
    'email': ['[email protected]'],
    'email_on_failure': False,
    'email_on_retry': False,
    'retries': 1,
    'retry_interval': timedelta(minutes=5),
    # 'queue': 'bash_queue',
    # 'pool': 'backfill',
    # 'priority_weight': 10,
    # 'schedule_interval': timedelta(1),
    # 'end_date': datetime(2016, 1, 1),
}

But when you break it down, it becomes:

args = {
    'owner': 'airflow',
    'depends_on_past': False,
    'start_date': datetime(2015, 1, 1),
    'email': ['[email protected]',],
    'email_on_failure': True,
    'email_on_retry': True,
}

And again,

t1 = BashOperator(
    task_id='print_date',
    bash_command='date',
    dag=dag)

t2 = BashOperator(
    task_id='sleep',
    bash_command='sleep 5',
    dag=dag)

which becomes:

t1 = BashOperator(
    task_id='print_date',
    bash_command='date',
    dag=dag)

t2 = BashOperator(
    task_id='sleep',
    email_on_failure=False,
    bash_command='sleep 5',
    dag=dag)

I am less concerned with the fact that you changed the code to illustrate the arguments and parameters and more concerned that you say: "Here is an example of a basic pipeline definition. Do not worry if this looks complicated, a line by line explanation follows below." leading us to believe that it will be a true reproduction, which it is not. Please fix the code in the tutorial so that it is clearer. Thanks!

Installation : Extra Packages

Hi,

I see in your documentation some subpackage (http://pythonhosted.org/airflow/installation.html#extra-packages). But I never heard about that with pip, moreover I don't see them using pip search airflow.

Tutorial documentation: "date"?

Just following the tutorial to see what airflow is about. It was easy to follow up till the "testing" paragraph on the tutorial page. Suddenly a "date" parameter is required.

Until there, I assumed you could run a batch job on anything and just request it to be run. Here, airflow implicitly requires that you run any DAG for a date or date ranges and that this parameter is required. I'm not sure if this parameter is passed into every task, or whether airflow simply needs this for another purpose.

It would be good to explain what this strange date parameter is for, what impact it has on tasks or whether this is internal use.

Can you have a DAG dependent on another DAG?

I see airflow.operators.ExternalTaskSensor supports dependency on an ask in another DAG. But is it possible to set dependency on another DAG itself? In other words, can external_task_id parameter be optional?

Question about TaskInstance.is_queueable

Would you mind clarifying the following condition in TaskInstance.is_queueable:

if self.execution_date > datetime.now() - self.task.schedule_interval:

I would have expected:

if self.execution_date > datetime.now():

Consider the use case of having a daily task. Let's suppose the task ran early this morning for 2015-07-01. However, the task failed for some reason. After fixing the issue, I go to backfill using a command such as:

airflow backfill -s 2015-07-01 -e 2015-07-01 dag_id

As the code is currently written, the backfill will not execute, due to the conditional that I mention above. I could hack it to work by passing in yesterday's date to backfill and then using "{{ tomorrow_ds }}" in my tasks, but this seems like it shouldn't be necessary.

If you're okay with the above change, I'm happy to submit a PR. If I should be taking a different approach, please let me know. Thanks!

setup.py download URL

While installing I noticed the url and download_url point to @mistercrunch's repo:

https://github.com/airbnb/airflow/blob/9a791999b871d5f19550cbf882fbd09037d99e52/setup.py#L69

    url='https://github.com/mistercrunch/Airflow',
    download_url=(
        'https://github.com/mistercrunch/Airflow/tarball/' + version),

Is that the correct repository to install and update from?

Any plans to add an SQS hook and SQS Sensor?

I would add the following actions to the Sensor?

1. Check for new message
2. Check for empty queue
3. Check for non-empty queue

Issue 23 Explanation

I am re-reading the last sentence of your answer of this issue #23.

"If you're planning on doing some large distributed computation that uses lots of memory, Airflow might not be the right platform. Instead, you may want to use a distributed system, or write some sort of service, and invoke it from Airflow."

I am not sure to really understand. Could you give me more information about it?

Service for scheduler

Hi

I'm just trying to figure how the tasks are run in production. Am I right in thinking that the scheduler must be running, and if i am using celery than a worker must be running as well?

I just plan on using mysql at the moment. Do you recommend creating a init script to start the scheduler?

Cheers

Using templates in PythonOperator

Sorry for a trivial question, I've been trawling thru the code but cant seem to get templating to work within PythonOperators.

I've been using the templating features of the PostgresOperator to read an external sql file and template some of the logic. I'd like to do similar using a PythonOperator, specifically:

• Call a python function supplying a dict (via a PythonOperator)
• The python function should load a template and render it using the supplied dict as variables.

Is this possible? I've been fiddling trying to supply a template_fields value to my function hoping it will magically work.

I need a flow to start with a sensor

I have a flow that is linear and sequential as shown below :

1. new data arrives in SQS
2. wait for that data to start being written to postgres
3. wait for the SQS queue to drain
4. send an email that the flow completed

Eventually, we will have a pre-cursor step (e.g. step 0.) that kicks off a Spark job before waiting for new data in the SQS queue. For now, I do not have that, so I would like my flow to simply reset to the first step. In some ways, I am asking for a circular flow that never ends. This breaks the entire workflow abstraction. On another note, where is the "cron" functionality? Do I schedule airflow commands in cron outside of airflow?

Can you add a slightly more realistic example of a data pipeline in the cloud?

For example, the Luigi example walks through a case that involves importing data into a DB. It would be cool if there were some examples that read from one location (e.g. S3) and wrote to another (e.g. DB).

depends_on_past at DAG level

I have a DAG which looks like the following:
t1 -> t2

When I run a backfill for this DAG (say for day1 and day2) , I want to make sure that the DAG for day2 gets executed only after the DAG for day1 is complete. From the name, "depends_on_past" looked like the right thing to use. However, it works at the task level. In other words, t1 for day2 starts executing after t1 for day1 is complete before t2 for day1 is completed.

Is there a way to force a similar "depends_on_past" at the DAG level?

TimeSensor does not trigger

We are trying to kick off our pipeline by a TimeSensor, but it does not trigger. There is no error in the logs visible - the graph looks like this:

And the code like this:

from airflow import DAG
from airflow.operators import DummyOperator, MySqlOperator, BashOperator
from datetime import datetime, time, timedelta
from airflow.models import Variable
from airflow.operators.sensors import TimeSensor
from airflow.operators.sensors import SqlSensor

default_args = {
    'owner': 'deploy',
    'depends_on_past': False,
    'start_date': datetime(2015,06,24),
    'queue': 'default',
}

# the DAG
dag = DAG('pvi', default_args=default_args)

# tables
app_installs = SqlSensor(conn_id="exasol", sql="""SELECT MAX(CREATED_TS) FROM PVI.APP_INSTALLS WHERE CREATED_DATE='{{ ds }}'""", task_id='table_app_installs', dag=dag)
app_logins = SqlSensor(conn_id="exasol", sql="""SELECT MAX(CREATED_TS) FROM PVI.APP_LOGINS WHERE CREATED_DATE='{{ ds }}'""", task_id='table_app_logins', dag=dag)

# time dependencies
midnight_sensor = TimeSensor(target_time=time(15, 28), task_id='midnight', dag=dag)
start_aggregation = DummyOperator(task_id='start_aggregation', dag=dag)
start_aggregation.set_upstream(midnight_sensor)
start_aggregation.set_downstream([app_logins, app_installs])

We tried to trigger the chain by putting the target_time onto something 5 minutes in the future, the DAG gets refreshed every 60 seconds (default) but nothing happens. Worker, Scheduler, Flower and Webserver are on the same machine running Python 2.7 and FreeBSD 9.3 and we have a second machine running a celery worker which is listening on a different queue (q2). We have another DAG that works as expected on q2.

Security question

I want to use s3 and I see that airflow works well with aws (ec2, s3). But s3 mean that I will have data outside. What do you have for security in airflow?

How do you feel about putting hook configuration in airflow.cfg?

We distribute our configs via text files managed by puppet so it would be awkward to put that into the db for airflow. Would you accept a patch that looks for conn_id in settings.conf before checking the database?

Can airflow synchronize with a CI like jenkins or replace it ?

Support for Amazon Elastic Container Service?

Just wondering if there's any plan to support Amazon ECS? I saw some discussion about potentially using docker containers for memory management. ECS would be a nice way to abstract away other details like selecting which server should run a given task.

Does airflow handle memory management ?

I didn't any ui pictures about that.

Issue using the PostGresOperator

Specifically, what is expected with respect to the "conn_id" argument?

The code below successfully connects to Postgres. I then pass the db_conn object to the PostgresOperator and get an exception shown at the bottom

db_conn_string = "host='localhost' dbname='cousteau_dev' user='siddharth' password='secret'"
db_conn = psycopg2.connect(db_conn_string)
print "Successfully Connected to database\n    ->%s and conn=%s" % (db_conn_string, db_conn)

wait_for_new_data_in_db = PostgresOperator(
    task_id='wait_for_new_data_in_db',
    postgres_conn_id=db_conn,
    sql='select count(*) from receiver_domain_aggregate;',
    dag=dag)
wait_for_new_data_in_db.set_upstream(wait_for_first_sqs_message)

Here's the exception

2015-06-21 20:21:06,789 - root - INFO - Executing <Task(PostgresOperator): wait_for_new_data_in_db> for 2015-06-12 00:00:00
2015-06-21 20:21:06,801 - root - INFO - Executing: select count(*) from receiver_domain_aggregate;
2015-06-21 20:21:06,805 - root - ERROR - (psycopg2.ProgrammingError) can't adapt type 'psycopg2.extensions.connection' [SQL: 'SELECT count(*) AS count_1 \nFROM (SELECT connection.id AS connection_id, connection.conn_id AS connection_conn_id, connection.conn_type AS connection_conn_type, connection.host AS connection_host, connection.schema AS connection_schema, connection.login AS connection_login, connection.password AS connection_password, connection.port AS connection_port, connection.extra AS connection_extra \nFROM connection \nWHERE connection.conn_id = %(conn_id_1)s) AS anon_1'] [parameters: {'conn_id_1': <connection object at 0x10dabe640; dsn: 'host='localhost' dbname='cousteau_dev' user='siddharth' password=xxxxxxxx', closed: 0>}]
Traceback (most recent call last):
  File "/usr/local/lib/python2.7/site-packages/airflow/models.py", line 746, in run
    task_copy.execute(context=self.get_template_context())
  File "/usr/local/lib/python2.7/site-packages/airflow/operators/postgres_operator.py", line 36, in execute
    self.hook = PostgresHook(postgres_conn_id=self.postgres_conn_id)
  File "/usr/local/lib/python2.7/site-packages/airflow/hooks/postgres_hook.py", line 25, in __init__
    if db.count() == 0:
  File "/usr/local/lib/python2.7/site-packages/sqlalchemy/orm/query.py", line 2734, in count
    return self.from_self(col).scalar()
  File "/usr/local/lib/python2.7/site-packages/sqlalchemy/orm/query.py", line 2503, in scalar
    ret = self.one()
  File "/usr/local/lib/python2.7/site-packages/sqlalchemy/orm/query.py", line 2472, in one
    ret = list(self)
  File "/usr/local/lib/python2.7/site-packages/sqlalchemy/orm/query.py", line 2515, in __iter__
    return self._execute_and_instances(context)
  File "/usr/local/lib/python2.7/site-packages/sqlalchemy/orm/query.py", line 2530, in _execute_and_instances
    result = conn.execute(querycontext.statement, self._params)
  File "/usr/local/lib/python2.7/site-packages/sqlalchemy/engine/base.py", line 914, in execute
    return meth(self, multiparams, params)
  File "/usr/local/lib/python2.7/site-packages/sqlalchemy/sql/elements.py", line 323, in _execute_on_connection
    return connection._execute_clauseelement(self, multiparams, params)
  File "/usr/local/lib/python2.7/site-packages/sqlalchemy/engine/base.py", line 1010, in _execute_clauseelement
    compiled_sql, distilled_params
  File "/usr/local/lib/python2.7/site-packages/sqlalchemy/engine/base.py", line 1146, in _execute_context
    context)
  File "/usr/local/lib/python2.7/site-packages/sqlalchemy/engine/base.py", line 1341, in _handle_dbapi_exception
    exc_info
  File "/usr/local/lib/python2.7/site-packages/sqlalchemy/util/compat.py", line 199, in raise_from_cause
    reraise(type(exception), exception, tb=exc_tb)
  File "/usr/local/lib/python2.7/site-packages/sqlalchemy/engine/base.py", line 1139, in _execute_context
    context)
  File "/usr/local/lib/python2.7/site-packages/sqlalchemy/engine/default.py", line 450, in do_execute
    cursor.execute(statement, parameters)
ProgrammingError: (psycopg2.ProgrammingError) can't adapt type 'psycopg2.extensions.connection' [SQL: 'SELECT count(*) AS count_1 \nFROM (SELECT connection.id AS connection_id, connection.conn_id AS connection_conn_id, connection.conn_type AS connection_conn_type, connection.host AS connection_host, connection.schema AS connection_schema, connection.login AS connection_login, connection.password AS connection_password, connection.port AS connection_port, connection.extra AS connection_extra \nFROM connection \nWHERE connection.conn_id = %(conn_id_1)s) AS anon_1'] [parameters: {'conn_id_1': <connection object at 0x10dabe640; dsn: 'host='localhost' dbname='cousteau_dev' user='siddharth' password=xxxxxxxx', closed: 0>}]

Since I can connect using psycopg2, I'm going to revert to using a PythonOperator. My flow is entirely made up of PythonOperators. I found it easier to use.

For CeleryExecutor, why do we need to install mysql?

We don't use mysql, but there is a requirement to have the mysql client installed. We get the exception when starting the airflow worker.

How do you log print statements?

I'm running a few scripts via PythonOperator and was wondering how I can log print statements within these scripts. Any information would be appreciated. Thank you,

Kevin

Inter-dag deps

I'm trying to use the ExternalTaskSensor to execute a data processing dag upon the completion of another dag that builds some lookup tables.

My child dag's demo code is below:

random_dep = ExternalTaskSensor(
    external_dag_id='driver',
    external_task_id='templated',
    owner='crm',
    task_id='random_dep'
    )

run_this = PythonOperator(
    task_id='print_the_context',
    provide_context=True,
    python_callable=print_context,
    dag=dag)
run_this.set_upstream(random_dep)

task = PythonOperator(
    task_id='sleep_for_1',
    python_callable=my_sleeping_function,
    op_kwargs={'random_base': 1},
    dag=dag)

task.set_upstream(run_this)

Have a correctly used the ExternalTaskSensor? Ive noticed this dag isnt rendered on the "Graph View" of the frontend.

Supported Python versions

What are the supported Python versions? Are there plans to support Python 3?

Configuration file issue, plugin folder key name not specified in configuration file.

System details:

Description: Ubuntu 15.04
Linux my-pc 3.19.0-21-generic #21-Ubuntu SMP Sun Jun 14 18:31:11 UTC 2015 x86_64 x86_64 x86_64 GNU/Linux

Steps to reproduce:

I cloned the repo and tried the quick start steps : http://pythonhosted.org/airflow/start.html

got error at the following step:

$ airflow initdb
Traceback (most recent call last):
File "/usr/local/bin/airflow", line 3, in
from airflow.bin.cli import get_parser
File "/usr/local/lib/python2.7/dist-packages/airflow/init.py", line 10, in
from airflow.configuration import conf
File "/usr/local/lib/python2.7/dist-packages/airflow/configuration.py", line 276, in
conf.read(AIRFLOW_CONFIG)
File "/usr/lib/python2.7/ConfigParser.py", line 305, in read
self._read(fp, filename)
File "/usr/lib/python2.7/ConfigParser.py", line 546, in _read
raise e
ConfigParser.ParsingError: File contains parsing errors: /home/viki/airflow/airflow.cfg
[line 32]: '/home/viki/airflow/plugins\n'

The error seems to be because the configuration file is like this

Where your Airflow plugins are stored

/home/viki/airflow/plugins

i changed it to

Where your Airflow plugins are stored

plugins_folder = /home/viki/airflow/plugins

and it worked.

example_dags/cron_replacement.py

I really like the concept of Airflow but I've been stumped trying to get a simple cron replacement example DAG working.

I took the example_bash_operator.py and changed args like so:

 args = {
    'owner': 'airflow',
    'start_date': datetime.now(),
    'schedule_interval': timedelta(minutes = 5)
}

I was expecting this DAG to start now and run every 5 minutes, but instead the scheduler runs the DAG in a loop, back-to-back.

I'd imagine migrating from cron to Airflow would be a pretty common use case - it would be awesome to see an example DAG for replacing e.g. a cron that runs every 6 hours.

EmailOperator is not working on Mac with Postfix : expecting TLS

Why do I need TLS? This requires setting it up for any smtp server, even one I run locally.

2015-06-19 13:20:09,793 - root - INFO - Executing <Task(EmailOperator): email_pipeline_start> for 2010-10-10 00:00:00
2015-06-19 13:20:10,459 - root - ERROR - STARTTLS extension not supported by server.
Traceback (most recent call last):
  File "/usr/local/lib/python2.7/site-packages/airflow/models.py", line 746, in run
    task_copy.execute(context=self.get_template_context())
  File "/usr/local/lib/python2.7/site-packages/airflow/operators/email_operator.py", line 36, in execute
    send_email(self.to, self.subject, self.html_content)
  File "/usr/local/lib/python2.7/site-packages/airflow/utils.py", line 339, in send_email
    send_MIME_email(SMTP_MAIL_FROM, to, msg)
  File "/usr/local/lib/python2.7/site-packages/airflow/utils.py", line 349, in send_MIME_email
    s.starttls()
  File "/usr/local/Cellar/python/2.7.9/Frameworks/Python.framework/Versions/2.7/lib/python2.7/smtplib.py", line 644, in starttls
    raise SMTPException("STARTTLS extension not supported by server.")
SMTPException: STARTTLS extension not supported by server.
2015-06-19 13:20:10,503 - root - ERROR - Failed to send email to: ['[email protected]']
2015-06-19 13:20:10,503 - root - ERROR - STARTTLS extension not supported by server.
2015-06-19 13:20:10,503 - root - ERROR - STARTTLS extension not supported by server.
Traceback (most recent call last):
  File "/usr/local/bin/airflow", line 10, in <module>
    args.func(args)
  File "/usr/local/lib/python2.7/site-packages/airflow/bin/cli.py", line 195, in test
    ti.run(force=True, ignore_dependencies=True, test_mode=True)
  File "/usr/local/lib/python2.7/site-packages/airflow/models.py", line 749, in run
    raise e
smtplib.SMTPException: STARTTLS extension not supported by server.

I also tested this out with our production postfix server, and that didn't work for a different reason, which is not apparent yet. Which email SMTP servers have you tried this out with? Examples would be very helpful.

Integrity error: Data Profiling -> Charts

So I've added an external connection successfully. A simple ad-hoc query returns data as expected.

However when I try to add a chart I get the following error:

Integrity error. (sqlite3.IntegrityError) NOT NULL constraint failed: chart.conn_id

It seems as if the sqlite job database isn't able to register the conn_id of my new external source?

I am able to see the source that I added "redshift" in the charting screen.

Operating system:

Distributor ID: Ubuntu
Description:    Ubuntu 14.04.1 LTS
Release:        14.04
Codename:       trusty