Rename process.runtime.jvm.buffer.* metrics

[trask]

Rename process.runtime.jvm.buffer.* metrics to process.runtime.jvm.buffers.* based on guidance from https://github.com/open-telemetry/opentelemetry-specification/blob/main/specification/metrics/semantic_conventions/README.md#pluralization (also similar to system.processes.*).

Also suggesting a couple other renames because buffer "usage" and "limit" don't seem super clear that they're about memory. And also these proposed names match the underlying MBean attributes.

So in total, proposing:

• process.runtime.jvm.buffer.count --> process.runtime.jvm.buffers.count (maybe .active instead of .count, or maybe just dropping the .count, depending on outcome of other discussions)
• process.runtime.jvm.buffer.usage --> process.runtime.jvm.buffers.memory_usage
• process.runtime.jvm.buffer.limit --> process.runtime.jvm.buffers.total_capacity

Not adding this to JVM metric stability project since we are not planning to include these metrics in the initial stability.

[breedx-splk]

I like this. Just one small nitpicky bikeshedding tweak around process.runtime.jvm.buffers.total_capacity.

I think the former limit carries with it some idea about maximum capacity, or upper achievable bounds. On the other hand, total seems to imply some summarized aggregation and can lose the fact that it's a boundary. With that, I'm proposing:

• process.runtime.jvm.buffer.limit --> process.runtime.jvm.buffers.max_capacity