arthur-plautz opened a new pull request, #51254:
URL: https://github.com/apache/airflow/pull/51254
<!--
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. -->
---
The goal of this draft is to discuss the inclusion of two classes
representing different approaches on achieving Change Data Capture at some
level in the airflow context:
- **Table Row Count**: A simple approach for getting changes from an
external table without the overhead of previous configuration or exploration.
For some use cases, not having reliable columns may impose a challenge to many
CDC approaches, thus, the idea of capturing the row count presents a simpler
alternative that wouldn't require any prerequisites other than a working
connection with the database.
- **Logical Replication**: CDC at transaction level. This is a more robust
solution that requires a certain level of control over the target database and
the replication itself. The details and prerequisites for this approach are
detailed in the [Postgres Logical
Replication](https://www.postgresql.org/docs/current/logical-replication.html)
documentation, this approach assumes that the user has the knowledge and access
to perform the required steps to achieve database logical replication.
Since the operations related to these triggers are mostly I/O bound, using a
synchronous connection to periodically operate in the triggerer loop may cause
performance problems, specially when dealing with small `poke_interval` values.
This solution is presented here as the first go to option, observing the well
established use of the [Postgres
Hook](https://airflow.apache.org/docs/apache-airflow-providers-postgres/stable/_api/airflow/providers/postgres/hooks/postgres/index.html).
Alternatively, using the [asyncpg](https://github.com/MagicStack/asyncpg)
library may be an option, with the cost of introducing an additional dependency
and a different implementation on Postgres connection.
The decision on the state tracking mechanism required for both approaches is
still open to discussion, although the current intended design choice leans
towards the one proposed in the
[AIP-30](https://cwiki.apache.org/confluence/display/AIRFLOW/AIP-30%3A+State+persistence).
Tests, examples and documentation will be added soon.
A similar proposal is presented in the PR linked below, although the nature
of the changes is complementary to this work.
Related: [51056](https://github.com/apache/airflow/pull/51056)
--
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]
