Log delta count in addition to throughput in LoggingMeterRegistry

[fstaudt]

Closes #5548

[shakuzen]

Thanks for the pull request and quick action on review. This looks good to me. We'll have to wait to merge until we start 1.15 development on main.

[jonatan-ivanov]

What do you think about calling it delta instead of delta_count?

[fstaudt]

Thanks for the pull request and quick action on review. This looks good to me. We'll have to wait to merge until we start 1.15 development on main.

Thanks for your quick review.

I was hoping that this change could still be part of 1.14.0 GA.
What is the reason that it should wait for 1.15 ?

[fstaudt]

What do you think about calling it delta instead of delta_count?

I woud keep delta_count in the logs.
The internal field is named count, it feels right to keep count in the logs.
If we only keep delta, it would make it less clear for users that it is a indeed a count.

[shakuzen]

I was hoping that this change could still be part of 1.14.0 GA.
What is the reason that it should wait for 1.15 ?

Our general policy is that we do not merge enhancements after a release candidate is released. 1.14.0-RC1 was released yesterday, so the next opportunity for enhancements is 1.15.0-M1.

[fstaudt]

I've noticed a discrepancy in logs for counters between delta_count and throughput when baseUnit is set.
I've added an extra commit with updated unit tests to fix this.

[fstaudt]

Our general policy is that we do not merge enhancements after a release candidate is released. 1.14.0-RC1 was released yesterday, so the next opportunity for enhancements is 1.15.0-M1.

OK, clear.

Do you already have an estimation of release date for 1.15.0 GA ?
It would help me for internal communication.

[jonatan-ivanov]

We have a 6 months minor release cycle, based on our current release schedule, 1.15.0 should go GA around the second Monday of May (12th) but you might be able to use 1.15.0-M1 in January.

[shakuzen]

I've gone ahead and added tentative dates for 1.15 to our support page: https://micrometer.io/support/

[jonatan-ivanov]

Was changing

" throughput=" + print.rate(count)

to

" throughput=" + print.unitlessRate(count)

intentional?