[AndyNemmity]

The stable chart thing is so weird. Internally use we some abstractions, but I looks at stable charts and it requires so much time just to understand all of what's going on. Everything is a variable pointed to values, and you can't reason about any of it.

It seems like the hope is, just ignore it all, and the docs are good, and just follow them, but I don't live in any kind of world I can do that.

And the commits, and the direction of all of them seem to go more and more impossible to read conditionally rendered symbols.

I've had such a challenge understanding and using helm well enough. Small gotchas everywhere that can just eat up tons of time. This doesn't feel like the end state to me.

[markbnj]

> It seems like the hope is, just ignore it all, and the docs are good, and just follow them, but I don't live in any kind of world I can do that.

Yep, agreed, we've used very few charts from stable, and in some cases where we have we needed to fork and change them, which is its own special form of suck. The one I contributed was relatively straightforward: a deployment, service and a configMap to parameterize and mount the conf file in the container at start. Even so I found it a challenge to structure the yaml in such a way that the configuration could expose the full flexibility of the binary, and in the end I didn't come anywhere near that goal. You take something like a chart for elasticsearch or redis and its just so much more complicated than that.

[AndyNemmity]

Right, I'm in particular working on charts for ELK, and it's just a mess. I just took down all my data (in staging, so all good) due to a PVC. The charts won't update without deleting them when particular parts of the chart change, but if you delete them, you lose your PVC data.

So I find the note in an issue somewhere stating, this is.. intentional?.. and that of course you need some annotation that will change it.

Let alone the number of things like, xpack, plugins, the fact that java caches the DNS so endpoints don't work on logstash, on and on.

It seems like everyone is saying operators are going to be the magical way to solve this, but if anything it seems like one set of codified values, that don't address any of the complexity.

[markbnj]

You're using a statefulset? Here's a tip: you can delete a statefulset without deleting the pods with `kubectl delete statefulset mystatefulset --cascade=false`. The pods will remain running, but will no longer be managed by a controller. You can then alter and recreate the statefulset and as long as the selector still selects those pods the new statefulset will adopt them. If you then need to update the pods you can delete them one at a time without disturbing the persistent volume claims, and the controller will recreate them.