Existing helm chart fails after recent changes on helm chart

[1yuv]

Describe the bug
Using existing helm chart that works on 5.1 would fail on latest master after recent changes from PR #10758 .

To Reproduce
Steps to reproduce the behavior:

1. Extract latest helm values from any 5.1 instance
2. Try installing same helm value with latest cht-core checked out.

helm upgrade namespace /path/to/medic/cht-core/scripts/build/helm -n namespace -f /path/to/yuvraj/namespace-values.yaml

3. This fails with following error.

Error: UPGRADE FAILED: template: cht-chart/templates/sentinel/deployment.yaml:22:32: executing "cht-chart/templates/sentinel/deployment.yaml" at <.Values.sentinel.node_env>: nil pointer evaluating interface {}.node_env

Expected behavior
A helm chart that works already should work without any issue.

Environment

• Instance: Any
• Browser: Any
• Client platform: (eg: Windows, MacOS, Linux)
• App: helm
• Version: latest-master

cc: @mrjones-plip

[mrjones-plip]

so glad you caught this - thanks @1yuv !

[vikrantwiz02]

Hii @mrjones-plip, @1yuv, @andrablaj , @jkuester
As this issue got pop-up due to my PR #10758 that got merged.

I think that the cleanest way to handle this is to swap those nested calls out for Helm's built-in dig function (e.g., {{ dig "sentinel" "node_env" "production" .Values }}). This will safely traverse the map and fall back to the default without crashing.

I also know that the api templates also share this exact same vulnerability, so I can patch those up too and drop a quick regression test in validate-templates.sh so it doesn't happen again.

Let me know if I'm correct, If all things ok, then you can assign me.

[1yuv]

Hi @vikrantwiz02 , that sounds like a plan to fix the issue. This is currenlty causing issues to anyone trying to upgrade to 5.1 if they check out latest master. So, I am assigning you to work on this.