WEBVTT 00:04.410 --> 00:11.790 We have just executed multiple tests using coproducer performance testability and here on the screen 00:11.790 --> 00:20.070 you see actually limits of my computer and the limits of coffeecake studied as Loung to hear, but indeed, 00:20.370 --> 00:26.510 all one million messages we are successfully sent at those mentioned here right now. 00:26.520 --> 00:30.240 Let's try to start Kafka consumer performance test. 00:30.240 --> 00:33.360 And for that, you need to use another comment, this one. 00:33.630 --> 00:39.590 And here, let's come back to our Kafka Gloster and read the messages from Topic Burov. 00:39.810 --> 00:46.890 And let's read those ten thousand messages and see how fast this consumer will be able to consume all 00:46.890 --> 00:47.220 of them. 00:47.610 --> 00:50.020 So let's hope this comment go here. 00:50.040 --> 00:51.690 Let's open up new tab. 00:51.990 --> 00:53.130 Go to Kafka. 00:54.220 --> 01:01.600 And let's start consumption here and here are results of testing, let's make it a bit smaller because 01:01.600 --> 01:04.260 actually just single roll of information. 01:04.660 --> 01:07.440 Let me make it a bit smaller once again, like this. 01:07.870 --> 01:14.500 And again, we are still not able to read it clearly, but we could distinguish that in total, we have 01:14.500 --> 01:19.850 consumed nine megabytes at the rate 13 megabits per second. 01:19.870 --> 01:21.040 So here was this rate. 01:21.340 --> 01:27.460 And in total, we have consumed the ten thousand messages and afterwards you'll see some additional 01:27.580 --> 01:28.230 timings. 01:28.660 --> 01:35.100 Let's not try to consume more messages and let's increase messages quantity to one hundred thousand. 01:35.410 --> 01:41.230 So let's go to our file, increase it to one hundred thousand like this. 01:41.770 --> 01:44.440 Copy come and go here based at. 01:45.800 --> 01:50.000 And all of those marriages were successfully consumed just in one second. 01:50.030 --> 01:56.870 So here you see a difference in time here with staff time and heroes and time and throughput rate was 01:56.870 --> 01:59.750 around 100 megabits per second. 01:59.900 --> 02:02.270 And that is a great result for my computer. 02:02.720 --> 02:06.120 And finally, let's try to consume one million messages. 02:06.380 --> 02:11.080 Let's go here and adjust to this number two, one million copies. 02:12.100 --> 02:13.120 And based here. 02:16.930 --> 02:22.930 And here you see a really, really nice result, I have not even expected such results and one million 02:22.930 --> 02:31.300 messages were received just in the two seconds here you see a difference in time and the rate was around 02:31.390 --> 02:33.760 400 megabits per second. 02:33.910 --> 02:35.500 Really, really nice rate. 02:35.890 --> 02:41.110 OK, we're getting that in previous example, when we have read the details about consumer group, I 02:41.110 --> 02:47.920 have told you that when consumer is not able to catch up with a lot, there will be a lag value that 02:47.920 --> 02:54.400 is greater than zero when, again, consumer is not able to read as as producer produces messages. 02:54.670 --> 03:02.170 And in the next level, let's try to simulate this case and we will create a new topic called The Power 03:02.170 --> 03:02.710 of Two. 03:02.710 --> 03:06.510 And we will decrease quantity of partitions to just, let's say, three. 03:06.700 --> 03:13.300 And afterwards we will try to start there, the producer and try to produce many messages to the topic 03:13.300 --> 03:15.820 and to consume them by specific consumers. 03:16.030 --> 03:17.120 Let's try that next. 03:17.140 --> 03:17.610 Bye bye.