Aliaksandr Valialkin

@arturminchukov , thank you for the pull request!

> I think, it's better to change influx parsing API. Instead of having `string` - immutable object, it's better to pass `[]byte` to the parser. It allows to mutate []byte...

> For example, with a hook into where two values are in conflict (similar to the Merge method here: https://github.com/golang/go/issues/18802#issuecomment-1884012504), and iteration over all the values, we can support additional...

> The capacity of each storage node can be estimated using [Little’s Law](https://en.wikipedia.org/wiki/Little%27s_law) as the product of its average RPS and average request duration. We already have the metric, which...

@makasim , let's implement accepting OTEL metrics with [delta temporality](https://opentelemetry.io/docs/specs/otel/metrics/data-model/#temporality) at VictoriaMetrics and vmagent side - just accept counter deltas and store them as is without transforming them to counters,...

> This way will result into irregular data points and poor [stale detection for range queries](https://docs.victoriametrics.com/keyConcepts.html#range-query). Which will make user experience even worse. This isn't true. There is zero practical...

The `process_*` metrics are usually read once per scrape interval. The usual scrape interval is in the range 10s - 30s. There is no practical sense in optimizing the code,...

I doubt this change is needed at all - see [this comment](https://github.com/VictoriaMetrics/metrics/issues/85#issuecomment-3148649935).

I agree that the one millisecond shouldn't be subtracted from the `tr.MaxTimestamp` - if the caller needs to select samples on a particular day, then it is the caller's responsibility...

> Is there any reason why TYPE should not be provided at all? The `TYPE` and `HELP` comments aren't provided by `github.com/VictoriaMetrics/metrics` package by default because of the following reasons:...