Matt Bolt

Overall, removing so much functionality concerns me, as opencost is a public package that has dependents. Since we're now also providing our own flyweight model for the resources, any dependents...

> Just curious, why can't you leverage output structures? We could absolutely leverage a work-around method here without too much of a problem. In general, I've found that using the...

@r2k1 > I suspect this is to prevent accidental modifications, but how likely is that? I think there should be no reason to modify the k8s objects. First of all,...

@r2k1 I've been very frustrated with the lack of flexibility in this API for a long time, and I totally support any and all motivation to improve (or completely re-work...

@ajfoijsdoiafdgjoasld Custom cost values do not apply retroactively. They should apply from the moment the change is made going forward. Let us know if that is not the case.

@mattray I haven’t looked myself but this wouldn’t surprise me at all. I doubt anyone is modifying the result, but you can still end up with go map access panics...

@Syndlex A first pass at most of this work is complete. Some of the break wasn't as clean as we wanted it to be, but we will iterate: * The...

@Noahnut Thanks for the PR here - I agree that this is important, and I think your point is valid. What I'm concerned about is exposing 2 configuration values that...

> Another want to ask, `err == nil` is the expected error handling? > > https://github.com/opencost/opencost/blob/59fdcb5260e086da91cb4fb61db7ea40ca5aa346/modules/prometheus-source/pkg/prom/datasource.go#L146-L149 Yeah, so I think you're right that Victoria Metrics doesn't return this value. That's...