Kubernetes deployment for production

Hi,
I have a kubernetes cluster in production. I want my temporal service to run in that cluster. As mentioned in the reference documents, using helm charts is not recommended for production. How do I approach this?

there are multiple factors to this:

  • the way helm sets up temporal services is in a production quality way.
  • the dependencies included in the helm chart are not intended for production.
  • helm in general isn’t a tool we recommend people use to deploy production services

suggested approach is to use the helm chart as a way to generate manifests (helm template) for only the temporal services (use dependencies external to the chart) and then use another method for deploying those manifests.

2 Likes

Hi derek,
Could you share how you generate your k8s manifests using helm?
I tried the helm template cmd, but did not make it.
Here is my command:

helm chart> helm template temporal-test temporal --output-dir ./k8s-manifests

But I always get the error ‘failed to download “temporal” (hint: running helm repo update may help)’.
From cmd manual, I see the usage is:

helm template [NAME] [CHART] [Flags]

I am new to helm and temporal. I do not understand what the NAME and CHART exactly are. In my cmd above, I just use ‘temporal’ as CHART because it’s value of the name keyword in Chart.yaml, and ‘temporal-test’ is an arbitrary identifier given as NAME.
Any help would be appreciated.

In the helm charts repo you should be able to just run

helm template .

It has the structure already for running this command.