Skip to content

Conversation

@ansd
Copy link
Member

@ansd ansd commented Jun 6, 2024

Monotonic time instead of wall clock time should be used to measure a duration.

This is mostly a cosmetic change given that consumer timeouts are 30 mins by default.

See also
rabbitmq/khepri#239
#10928

Monotonic time instead of wall clock time should be used to measure a
duration.

This is mostly a cosmetic change given that consumer timeouts are 30
mins by default.

See also
rabbitmq/khepri#239
#10928
@ansd ansd merged commit f5d2fd6 into main Jun 6, 2024
@ansd ansd deleted the monotonic-time branch June 6, 2024 14:34
@michaelklishin
Copy link
Collaborator

@Mergifyio backport v3.13.x

@mergify
Copy link

mergify bot commented Jun 6, 2024

backport v3.13.x

✅ Backports have been created

@ansd
Copy link
Member Author

ansd commented Jun 6, 2024

erlang:monotonic_time/1 is slower compared to os:system_time/1.
However this slowness is not noticeable with the following workload:

make run-broker FULL=1 RABBITMQ_SERVER_ADDITIONAL_ERL_ARGS="+S 1"

java -jar target/perf-test.jar --predeclared --exchange amq.default -k cq -u cq -A 10000 --qos 10000 -z 90

michaelklishin added a commit that referenced this pull request Jun 6, 2024
Prefer monotonic time to measure durations (backport #11389)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants