[go: up one dir, main page]

Page MenuHomePhabricator

Deploy airflow scheduler via helm chart
Closed, ResolvedPublic

Assigned To
Authored By
bking
Jun 28 2024, 2:54 PM
Referenced Files
F57464654: image.png
Sep 5 2024, 4:00 PM
F57464652: image.png
Sep 5 2024, 4:00 PM
F57464649: image.png
Sep 5 2024, 4:00 PM
F57464647: image.png
Sep 5 2024, 4:00 PM

Description

Once this patch is completed/merged, we'll have the airflow webserver running. The scheduler is another requirement, so this ticket is dedicated to adding the scheduler to our helm release. See the Airflow: High Availability Strategy doc for more details on this process.

Event Timeline

Gehel triaged this task as High priority.Jul 1 2024, 6:42 PM

Change #1070245 had a related patch set uploaded (by Brouberol; author: Brouberol):

[operations/deployment-charts@master] airflow: fully generate airflow.cfg from helm values

https://gerrit.wikimedia.org/r/1070245

Change #1070245 merged by Brouberol:

[operations/deployment-charts@master] airflow: fully generate airflow.cfg from helm values

https://gerrit.wikimedia.org/r/1070245

Change #1070255 had a related patch set uploaded (by Brouberol; author: Brouberol):

[operations/puppet@production] global_config: define an external-services entry for mx[1-2]001.wikimedia.org

https://gerrit.wikimedia.org/r/1070255

Change #1070255 merged by Brouberol:

[operations/puppet@production] global_config: define an external-services entry for mx[1-2]001.wikimedia.org

https://gerrit.wikimedia.org/r/1070255

Change #1070619 had a related patch set uploaded (by Brouberol; author: Brouberol):

[operations/deployment-charts@master] airflow: deploy the scheduler as a sidecar container

https://gerrit.wikimedia.org/r/1070619

I've been able to deploy the Airflow scheduler to our airflow-test-k8s instance! Everything looks nominal.

NAME                                READY   STATUS    RESTARTS   AGE
airflow-scheduler-974c47bcf-8lwj5   1/1     Running   0          104s
airflow-webserver-b6dc74bdd-htcn5   2/2     Running   0          104s

We can see our (single, for now) DAG: addition, that computes ... 1 + 2.

image.png (898×2 px, 163 KB)

I've been able to manually trigger it, and it ran successfully.

image.png (2×2 px, 294 KB)

The scheduler was able to serialize the dag and store it to DB, so that the webserver is able to display its code.

image.png (2×2 px, 460 KB)

All required components are running (the triggerer is optional, and we don't run it):

image.png (550×772 px, 44 KB)

brouberol opened https://gitlab.wikimedia.org/repos/data-engineering/airflow-dags/-/merge_requests/821

test-k8s: start regularly scheduling DAGs now that we run the scheduler

brouberol merged https://gitlab.wikimedia.org/repos/data-engineering/airflow-dags/-/merge_requests/821

test-k8s: start regularly scheduling DAGs now that we run the scheduler

Change #1070619 merged by Brouberol:

[operations/deployment-charts@master] airflow: deploy the scheduler via a separate Deployment

https://gerrit.wikimedia.org/r/1070619

Change #1071077 had a related patch set uploaded (by Brouberol; author: Brouberol):

[operations/deployment-charts@master] airflow: fix badly formatted Deployment separation

https://gerrit.wikimedia.org/r/1071077

Change #1071077 merged by Brouberol:

[operations/deployment-charts@master] airflow: fix badly formatted Deployment separation

https://gerrit.wikimedia.org/r/1071077

Change #1071153 had a related patch set uploaded (by Brouberol; author: Brouberol):

[operations/deployment-charts@master] airflow: enable visualizing logs of DAG runs in the webserver UI

https://gerrit.wikimedia.org/r/1071153

Change #1071153 merged by jenkins-bot:

[operations/deployment-charts@master] airflow: enable visualizing logs of DAG runs in the webserver UI

https://gerrit.wikimedia.org/r/1071153