Our team launched Astronomer v0.3.0 two weeks ago, timely delivering our strongest release yet. We set a solid foundation for both our Enterprise and Cloud Editions and are getting in the rhythm of shipping releases regularly and openly. In that spirit, read below for a short update that breaks down v.0.3.2, officially released this week.
A Breakdown of v0.3.2
1. All charts and their pods are now selected by a NetworkPolicy that defines strict Ingress rules
This will prevent pods from talking to pods they shouldn’t be talking to. Individual Airflow deployments will now also be isolated at the network level. We’re looking to define Egress rules in the future, but these improvements give us a good start.
2. Prevent redis and flower from restarting during code deployments
This dramatically upgrades the deployment sequence/experience. Flower now reliably remains up and shows workers going up or down in real time, making deployments significantly cleaner. You’ll find that workers are able to properly stop taking tasks, drain ongoing tasks, shut down, and restart.
3. Improved docker caching
This speeds up the development experience as well as new code deployment. It’s a huge improvement, and we’d like to potentially use a multi-stage onbuild image to optimize it further in the future.
4. A few things we now support
- The ability to run multiple Astronomers on a cluster
- EKS
- Google Cloud SQL
5. Updates to Prometheus/Grafana
To support multiple Airflow deployments, we added more descriptive labels and fixed network metrics that are all scoped to a particular namespace.
Join Our Program
Interested in trying this out yourself? Reserve a spot in our program, and we’ll reach out to schedule a demo.
