Some easier way to deploy temporal cluster in Kubernates. May be some direct yaml files?

Community Support Server Deployment
1
  1. They recently removed the K8S deployment part from the docker-compose repo. So no one can find it easily.
  2. You won’t find a single document which explain clearly about how you can deploy it in kubernates.
  3. You will find everything about temporal except its deployment to production
  4. I created most of my application according to temporal but now I don’t have a way to deploy it in prod.
2

I have been here at Temporal for 3 years now and have been very active in the community as well as helping cloud customers. Can tell you with outmost confidence that what you are claiming here as far as title goes especially is absolutely far from the truth and not correct. Sorry that you feel this way.

One of the things Temporal as a company imo values the most is user support. We go (and I know because have been part of this effort for a long time) above and beyond to do anything we can so any user is successful with Temporal.

Do admit that our kubernetes deployment resources and especially helm-chart repo is outdated and needs more attention. We are working hard on making sure this changes. It would also really help to get help from community as I a know that there is a lot of users running Temporal on very large scale on k8s and have been for a long time.

| 1. You will find everything about temporal except its deployment to production
Docs have a self-hosted guide Self-hosted Temporal Cluster guide | Temporal Documentation
and we provide free help on any deployment questions on our community Slack

| 1. I created most of my application according to temporal but now I don’t have a way to deploy it in prod.
What are the issues you are running to (please create a separate thread for concrete questions if possible).

I think we all thank you for the post and even if I personally do not agree with the title, we take it as good input to improve. Think that understanding the issues that you ran and are running into in more details would be very useful.

3

Installation is not “plug and play” but I would say the documentation and resources have improved since I first experimented with self-hosting about a year ago.

I used the helm chart for testing on Kubernetes. It isn’t production ready, but if you are going to be running it in production you would do yourself a favour getting to know how it all works so that you can support Temporal once you deploy it. Reading and understanding Self-hosted Temporal Cluster guide | Temporal Documentation will pay you back in the long run.

For my production deployment I used helm to render the charts out to local copies without all the dependencies (Postgres, Elasticsearch, etc.):

  • Visibility store is Postgres (not deployed on k8s) to get support from our DBA team
  • Elasticsearch advanced visibility store is hosted on k8s (but deployed using the vendor’s operator to ensure it meets best practices)

I used kustomize with the rendered helm charts to provide last mile configuration for each of our DEV, QA and PROD clusters; tweaking replica counts for each temporal service; integration with the pre-existing Prometheus and Grafana deployments; AAD integration for the Web UI; etc.

You may find these references useful in helping your understanding: