WEBVTT

00:04.230 --> 00:10.470
We have just started to multiple producers in Gloucester, created the new topical Burov, and also

00:10.470 --> 00:16.680
we have Loung took off coproducer performance test with total quantity of recall's equal to one thousand,

00:16.680 --> 00:20.640
and they were sent at the rate one hundred messages per second.

00:20.850 --> 00:24.840
And size of every message was one kilobyte with no increase.

00:24.840 --> 00:32.340
A bit of those numbers and let's send the, let's say, ten thousand messages at the rate one thousand

00:32.340 --> 00:33.120
per second.

00:33.150 --> 00:37.140
So just add here zeros, inam records and throughput.

00:37.590 --> 00:44.670
Let's Akabi again this command go here clear terminal and start over again.

00:50.050 --> 00:57.120
Knowing this log records, we see much higher throughput rate around one megabyte per second, and in

00:57.130 --> 01:04.770
total, ten thousand records were successfully sent and the consumer was able to get all of those messages,

01:05.050 --> 01:10.480
listen to increase numbers a bit more and listen, I'll produce, let's say, one hundred thousand messages

01:10.480 --> 01:13.730
at throughput to, let's say ten thousand messages per second.

01:14.080 --> 01:21.190
Let's go to all file and adjust knowm records to one hundred thousand like this throughput will be ten

01:21.190 --> 01:21.700
thousand.

01:21.700 --> 01:25.270
And the let's see how it will handle this load.

01:25.660 --> 01:26.100
OK.

01:26.770 --> 01:28.210
Clear terminal based.

01:33.530 --> 01:39.170
And this test was successful as well, and on average, you see here ten thousand records per second

01:39.170 --> 01:46.040
at the throughput rate around 10 megabits per second, let's make it even more and less.

01:46.040 --> 01:51.720
Try to produce one million records at throughput 100 messages per second.

01:51.890 --> 01:52.640
Let's try it.

01:52.970 --> 01:56.270
I'm not sure whether my computer will be able to survive this long.

01:56.290 --> 02:00.020
But anyway, so one million records of throughput.

02:00.020 --> 02:02.410
One hundred thousand messages per second.

02:03.020 --> 02:04.250
Let's copy this command.

02:04.250 --> 02:05.030
Go here.

02:05.330 --> 02:07.160
Clear time, no wasted.

02:07.170 --> 02:08.090
Fingers crossed.

02:08.300 --> 02:08.900
Enter.

02:11.750 --> 02:18.470
And it seems that I have hit the limits of my computer and the maximum achieved rate was seventy three

02:18.470 --> 02:23.690
thousand records per second at the rate around 70 megabits per second.

02:24.080 --> 02:31.330
And you may notice that latency here was hugely increased in comparison with previous outputs.

02:31.670 --> 02:38.390
And that's what happens when you try to produce one million messages at the rate one hundred thousand

02:38.660 --> 02:39.910
messages per second.

02:40.100 --> 02:44.360
That's how you're able to use the coproducer performance built in test.

02:44.870 --> 02:51.470
Lets know also try to launch a consumer performance testing and let's do the next buy by.