amoghrajesh opened a new pull request, #71701:
URL: https://github.com/apache/airflow/pull/71701
<!-- SPDX-License-Identifier: Apache-2.0
https://www.apache.org/licenses/LICENSE-2.0 -->
<!--
Thank you for contributing!
Please provide above a brief description of the changes made in this pull
request.
Write a good git commit message following this guide:
https://chris.beams.io/posts/git-commit/
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.
For user-facing UI changes, please attach before/after screenshots (or a
short
screen recording) so reviewers can assess the visual impact.
Feel free to ping (in general) for the review if you do not see reaction for
a few days
(72 Hours is the minimum reaction time you can expect from volunteers) - we
sometimes miss notifications.
In case of an existing issue, reference it using one of the following:
* closes: #ISSUE
* related: #ISSUE
-->
---
##### Was generative AI tooling used to co-author this PR?
<!--
If generative AI tooling has been used in the process of authoring this PR,
please
change below checkbox to `[X]` followed by the name of the tool, uncomment
the "Generated-by".
-->
- [ ] Yes (please specify the tool below)
<!--
Generated-by: [Tool Name] following [the
guidelines](https://github.com/apache/airflow/blob/main/contributing-docs/05_pull_requests.rst#gen-ai-assisted-contributions)
-->
## What and Why
When `VaultBackend` is configured as a secrets backend for airflow, fetching
one secret from Vault costs four HTTP requests today. Only one of them reads
the secret.
| Request | Source |
|---|---|
| `POST auth/approle/login` | `_VaultClient._client` |
| `GET auth/token/lookup-self` | post-login check in `_client` |
| `GET auth/token/lookup-self` | staleness check in the `client` property |
| `GET <mount>/data/<path>` | the read |
Neither validation is local. `hvac.Client.is_authenticated()` calls
`lookup_token()`, which issues `auth/token/lookup-self` over the network every
time it is asked.
The first check runs immediately after a successful login, against a token
whose login response has already been received. The second guards a cached
client against expiry, which is reasonable in isolation, but the secrets
backend is rebuilt on every lookup so that client never lives long enough to
need guarding.
This replaces both with reactive validation instead. Issue the request, and
if Vault rejects the token with a 403, authenticate again and retry once. Vault
answers 403 for both an expired token and an unreadable path, so a single retry
is the only way to tell them apart without paying a round trip before every
request.
## Improvements
Set up airflow with a running hashicorp vault docker image by confuguring
these envs:
```shell
AIRFLOW__SECRETS__BACKEND=airflow.providers.hashicorp.secrets.vault.VaultBackend
AIRFLOW__SECRETS__BACKEND_KWARGS='{"url":"http://vault-test:8200","auth_type":"approle","role_id":"blahblah","secret_id":"blahblah","mount_point":"secret","variables_path":"variables","connections_path":null,"config_path":null}'
```
Enabled audit logs on vault using: `vault audit enable file
file_path=/vault/logs/audit.log`
Placed two variables on vault using:
```shell
vault kv put -mount=secret variables/demo value=hello
vault kv put -mount=secret variables/demo2 value=world
```
### One variable lookup
Before and after changes run on `airflow variables get demo`:
Before:
```shell
amoghrajesh.desai …/airflow vault-login-cost-reduction $!?
orbstack ♥ 12:30 docker exec vault-test cat /vault/logs/audit.log \
| jq -r 'select(.type=="request") | .request.path' | sort | uniq -c
1 auth/approle/login
2 auth/token/lookup-self
1 secret/data/variables/demo
```
After:
```shell
amoghrajesh.desai …/airflow vault-login-cost-reduction $!?
orbstack ♥ 12:26 docker exec vault-test cat /vault/logs/audit.log \
| jq -r 'select(.type=="request") | .request.path' | sort | uniq -c
1 auth/approle/login
1 secret/data/variables/demo
```
### Two variable lookup
Before and after changes, ran: `airflow variables get demo` and then
`airflow variables get demo2` in same session
Before:
```shell
amoghrajesh.desai …/airflow vault-login-cost-reduction $!?
orbstack ♥ 12:31 docker exec vault-test cat /vault/logs/audit.log \
| jq -r 'select(.type=="request") | .request.path' | sort | uniq -c
2 auth/approle/login
4 auth/token/lookup-self
1 secret/data/variables/demo
1 secret/data/variables/demo2
```
After:
```shell
amoghrajesh.desai …/airflow vault-login-cost-reduction $!?
orbstack ♥ 12:32 docker exec vault-test cat /vault/logs/audit.log \
| jq -r 'select(.type=="request") | .request.path' | sort | uniq -c
2
auth/approle/login
1 secret/data/variables/demo
1 secret/data/variables/demo2
```
### Summary
<img width="589" height="135" alt="image"
src="https://github.com/user-attachments/assets/5cb77e03-d7ce-46c2-9454-fdac93a8015c";
/>
**Scales linearly. Two variables fetched in one process: 8 requests before,
4 after.**
## Behaviour / Backcompat
`VaultHook.get_conn()` hands the raw client to callers who then issue their
own requests, so they cannot benefit from the retry the wrapped methods use. It
keeps validating and rebuilding exactly as before, which preserves transparent
re-authentication for that public entry point. The secrets backend never
goes through `get_conn()`, so its hot path stays probe free and gets the full
reduction.
---
* Read the **[Pull Request
Guidelines](https://github.com/apache/airflow/blob/main/contributing-docs/05_pull_requests.rst#pull-request-guidelines)**
for more information. Note: commit author/co-author name and email in commits
become permanently public when merged.
* For fundamental code changes, an Airflow Improvement Proposal
([AIP](https://cwiki.apache.org/confluence/display/AIRFLOW/Airflow+Improvement+Proposals))
is needed.
* When adding dependency, check compliance with the [ASF 3rd Party License
Policy](https://www.apache.org/legal/resolved.html#category-x).
* For significant user-facing changes create newsfragment:
`{pr_number}.significant.rst`, in
[airflow-core/newsfragments](https://github.com/apache/airflow/tree/main/airflow-core/newsfragments).
You can add this file in a follow-up commit after the PR is created so you
know the PR number.
--
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]
