JMeter Summariser: Flushed statistics according to interval even if no sample occured

[asfimport]

Liu XP (Bug 63411):
I have set summariser.interval=10 in jmeter.properties file. But it couldn't be flushed logs every time.

The following is part of detail logs. Please take a look at 2019/04/03 22:49:44 and 2019/04/03 22:52:35 timestamp.

...
2019/04/03 22:49:01 INFO - jmeter.reporters.Summariser: summary + 24582 in 00:00:10 = 2395.7/s Avg: 17 Min: 0 Max: 705 Err: 1 (0.0041%) Active: 64 Started: 45 Finished: 0
2019/04/03 22:49:01 INFO - jmeter.reporters.Summariser: summary = 13118986 in 00:59:39 = 3665.8/s Avg: 9 Min: 0 Max: 694682 Err: 10091 (0.0769%)
2019/04/03 22:49:11 INFO - jmeter.reporters.Summariser: summary + 23996 in 00:00:10 = 2449.8/s Avg: 17 Min: 0 Max: 747 Err: 0 (0.0000%) Active: 64 Started: 45 Finished: 0
2019/04/03 22:49:11 INFO - jmeter.reporters.Summariser: summary = 13142982 in 00:59:49 = 3662.5/s Avg: 9 Min: 0 Max: 694682 Err: 10091 (0.0768%)
2019/04/03 22:49:21 INFO - jmeter.reporters.Summariser: summary + 23835 in 00:00:10 = 2395.7/s Avg: 18 Min: 0 Max: 780 Err: 2 (0.0084%) Active: 64 Started: 45 Finished: 0
2019/04/03 22:49:21 INFO - jmeter.reporters.Summariser: summary = 13166817 in 00:59:59 = 3659.0/s Avg: 9 Min: 0 Max: 694682 Err: 10093 (0.0767%)
2019/04/03 22:49:22 INFO - jmeter.JMeter: Finished remote host: 10.100.17.231 (1554302962982)
2019/04/03 22:49:23 INFO - jmeter.JMeter: Finished remote host: 10.100.17.232 (1554302963128)
2019/04/03 22:49:23 INFO - jmeter.JMeter: Finished remote host: 10.100.17.234 (1554302963373)
2019/04/03 22:49:23 INFO - jmeter.JMeter: Finished remote host: 10.100.17.227 (1554302963662)
2019/04/03 22:49:23 INFO - jmeter.JMeter: Finished remote host: 10.100.17.228 (1554302963824)
2019/04/03 22:49:23 INFO - jmeter.JMeter: Finished remote host: 10.100.17.229 (1554302963846)
2019/04/03 22:49:44 INFO - jmeter.reporters.Summariser: summary + 5936 in 00:00:24 = 248.9/s Avg: 19 Min: 0 Max: 710 Err: 0 (0.0000%) Active: 2 Started: 45 Finished: 62
2019/04/03 22:49:44 INFO - jmeter.reporters.Summariser: summary = 13172753 in 01:00:22 = 3636.5/s Avg: 9 Min: 0 Max: 694682 Err: 10093 (0.0766%)
2019/04/03 22:49:44 INFO - jmeter.JMeter: Finished remote host: 10.100.17.233 (1554302984986)
2019/04/03 22:52:35 INFO - jmeter.reporters.Summariser: summary + 297 in 00:02:50 = 1.7/s Avg: 891 Min: 0 Max: 258593 Err: 0 (0.0000%) Active: 1 Started: 45 Finished: 63
2019/04/03 22:52:35 INFO - jmeter.reporters.Summariser: summary = 13173050 in 01:03:13 = 3473.4/s Avg: 9 Min: 0 Max: 694682 Err: 10093 (0.0766%) 2019-04-03 22:52:35
2019/04/03 22:52:35 INFO - jmeter.reporters.Summariser: summary + 253 in 00:00:00 = 2480.4/s Avg: 1045 Min: 0 Max: 258795 Err: 0 (0.0000%) Active: 0 Started: 45 Finished: 64
2019/04/03 22:52:35 INFO - jmeter.reporters.Summariser: summary = 13173303 in 01:03:13 = 3473.4/s Avg: 9 Min: 0 Max: 694682 Err: 10093 (0.0766%)
2019/04/03 22:52:35 INFO - jmeter.JMeter: Finished remote host: 10.100.17.230 (1554303155273)
2019/04/03 22:52:40 INFO - jmeter.JMeter: Interrupting RMI Reaper

I found when there was no sampler occurred, the entry point of interval in src/core/org/apache/jmeter/reporters/Summariser.java line 174 would not be executed.

OS: All

[asfimport]

@pmouawad (migrated from Bugzilla):
Hello,
Thanks for report.
Yes you are right, in fact the 10s is in reality "10 s if there were SampleResults" received in interval.

If you have a better proposal, feel free to contribute a PR or some description of what you would prefer.

Thanks
Regards

[asfimport]

Liu XP (migrated from Bugzilla):
Can we add a message line with empty result data as below?

2019/04/03 22:49:54 INFO - jmeter.reporters.Summariser: summary + 0 in 00:00:34 = 0/s Avg: 0 Min: 0 Max: 0 Err: 0 (0.0000%) Active: 2 Started: 45 Finished: 62
2019/04/03 22:49:54 INFO - jmeter.reporters.Summariser: summary = 13172753 in 01:00:34 = 3536.5/s Avg: 9 Min: 0 Max: 694682 Err: 10093 (0.0766%)

[asfimport]

@pmouawad (migrated from Bugzilla):
The flushing would need to be changed, since currently it only occurs on sampleOccurred(SampleEvent e), so no sample == no flushing.

We would need to decouple results collection in sampleOccurred(SampleEvent e) from flushing which should then be controlled by a timer.

Would you be willing to propose a PR ?

Thanks

[asfimport]

Liu XP (migrated from Bugzilla):
Hi Philippe,

Where can I propose a PR?

[asfimport]

@pmouawad (migrated from Bugzilla):
(In reply to Liu XP from comment 4)

Hi Philippe,

Where can I propose a PR?

Hello Liu (I hope I don't make a mistake on your first name),

You can propose a PR here:

• https://github.com/apache/jmeter

Read this for all details:

• https://github.com/apache/jmeter/blob/trunk/CONTRIBUTING.md

Thanks for your proposals.

Regards