We change the main exporter to be stored and fetched using atomics rather
than aquiring a mutex for a mild (but significant in the disabled case) speedup.
Also has the benefit of not holding a global lock over all telemetry operations.
benchstat of logging benchmatks before and after:
name old time/op new time/op delta
Baseline-8 329ns ± 2% 327ns ± 2% ~ (p=0.181 n=19+17)
LoggingNoExporter-8 3.08µs ± 3% 2.42µs ± 2% -21.42% (p=0.000 n=20+19)
Logging-8 13.7µs ± 2% 13.2µs ± 1% -3.49% (p=0.000 n=19+19)
LoggingStdlib-8 5.39µs ± 3% 5.41µs ± 2% ~ (p=0.177 n=19+20)
This is a replacement for https://go-review.googlesource.com/c/tools/+/212244
but built on the single exporter principle rather than the exporter list.
Change-Id: Icc99319c4357e0bcb63386c64372a733e8a76796
Reviewed-on: https://go-review.googlesource.com/c/tools/+/221218
Run-TryBot: Emmanuel Odeke <emm.odeke@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
It is never invoked anyway, and it provices no useful benefit while complicating
the implementation, as it is the only non context aware method.
Change-Id: Id5a99439fedafdf4d71285e36103b4854cf3635a
Reviewed-on: https://go-review.googlesource.com/c/tools/+/221540
Run-TryBot: Ian Cottrell <iancottrell@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
Instead we only have a single exporter, and it must delegate behaviour
to any other exporters it wants to include.
This removes a whole collection of suprises caused by init functions adding
new exporters to a list, as well as generally making things faster, at the small
expense of needing to implement a custom exporter if you want to combine the
features of a few other exporters.
This is essentially the opposite of https://go-review.googlesource.com/c/tools/+/212243
which will now be abandoned in favor of this approach.
Change-Id: Icacb4c1f0f40f99ddd1d82c73d4f25a3486e56ce
Reviewed-on: https://go-review.googlesource.com/c/tools/+/220857
Run-TryBot: Ian Cottrell <iancottrell@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
Reviewed-by: Robert Findley <rfindley@google.com>
Make all the benchmarks cleanly pass around the context to remove a non logging
difference.
Rename the non logging calls benchmark to Baseline
Change-Id: I24a33b0a817df403fb43c53b5f097bc1e9418af4
Reviewed-on: https://go-review.googlesource.com/c/tools/+/220520
Run-TryBot: Ian Cottrell <iancottrell@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
There was a fragment of a sentence that must have been from a previous version
(as it talks about a return value for a function that does not have one).
Change-Id: I9d154fe10711344f93e1d49b68a811dbc9772710
Reviewed-on: https://go-review.googlesource.com/c/tools/+/212241
Run-TryBot: Ian Cottrell <iancottrell@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
This changes to use a mutex and directly execute the less performance
sensitive telemetry calls (tracing and logging) and then uses a submission
queue only for stats adjustments as those are much more sensitive (but it
should also be easier to keep up with them in bursts)
Fixesgolang/go#33692
Change-Id: Ia59a8975f21dfbfcf115be1f1d11b097be8dd9c8
Reviewed-on: https://go-review.googlesource.com/c/tools/+/190737
Run-TryBot: Ian Cottrell <iancottrell@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rebecca Stambler <rstambler@golang.org>
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
this shuffles things so there a single exporter API rather than an observer
It also removes most of the globals.
per telemetry type.
Change-Id: Iaa82abe2ded1fff9df8e067ed4a55bcbd9d9591f
Reviewed-on: https://go-review.googlesource.com/c/tools/+/190405
Run-TryBot: Ian Cottrell <iancottrell@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Emmanuel Odeke <emm.odeke@gmail.com>
