I'm working on a project that involves 10 microservices, and currently, I'm managing them all through a single values file in my Helm chart. Ideally, I'd like to have individual values files for each microservice to make it simpler to track changes. However, I need these services to be deployed under the same release name. My main issue arises during the upgrade process: when I run "helm upgrade," it seems to delete the old pods and only installs the pods for the last value file. I've looked into using helmfile, which can handle both installation and upgrades, but I would still need to write scripts to merge the values files each time I add a new service. Is there a way to automate this process so that I can simply add a new value file path without extra scripting?
3 Answers
You could also configure your Helm chart to include these value files directly in your chart's configuration (like in the `Chart.yaml`) so they get pulled in automatically. It could look something like this:
```
helm:
valueFiles:
- values/service1.yaml
- values/service2.yaml
- ...
```
However, I’d suggest being clear on the structure to make sure it works as intended.
Another approach would be to create separate values files for each microservice, like `microservice1.yaml`, `microservice2.yaml`, etc. Then you can deploy each one with a command like: `helm install microservice1 mychart -f microservice1.yaml`. If that's not working for your upgrades, it might be worth checking how you're managing the deployments in the same namespace.
I did that, but when I upgrade the values for the second microservice, it ends up deleting the previous one.
You can actually pass multiple values files when you install your Helm release by specifying the `-f` flag several times. Just make sure to include the `--reuse-values` flag during upgrades to preserve the existing values. This should help avoid deleting old pods. If it doesn't work, you might need to check if your value files are conflicting with each other since they may be defining different images.
I tried this too, but unfortunately, it still resulted in the old pods being deleted. Each value file has different Docker images.
Thanks for the tip! I'll give this a shot and see if it makes a difference.
Can you elaborate a bit more on how that works? I'm not entirely clear on the setup.