1
00:00:00,000 --> 00:00:02,780
‫Now let's talk about Amazon CodeGuru.

2
00:00:02,780 --> 00:00:05,117
‫And this is a machine learning-powered service

3
00:00:05,117 --> 00:00:06,800
‫that will do two things.

4
00:00:06,800 --> 00:00:09,090
‫Number one is automated code reviews,

5
00:00:09,090 --> 00:00:11,720
‫and number two is application performance recommendations.

6
00:00:11,720 --> 00:00:14,050
‫So when developers push our code,

7
00:00:14,050 --> 00:00:16,331
‫there is usually another developer that does a code review.

8
00:00:16,331 --> 00:00:18,865
‫And then when the code is deployed into production,

9
00:00:18,865 --> 00:00:21,660
‫you need to be able to monitor the performance of your code,

10
00:00:21,660 --> 00:00:24,180
‫and maybe you'll detect bugs by looking at the performance.

11
00:00:24,180 --> 00:00:26,491
‫So CodeGuru does that in an automated fashion.

12
00:00:26,491 --> 00:00:31,030
‫So CodeGuru Reviewer is here to do automated code reviews

13
00:00:31,030 --> 00:00:32,790
‫with static code analysis.

14
00:00:32,790 --> 00:00:36,250
‫So that means when you deploy your code onto a repository,

15
00:00:36,250 --> 00:00:38,497
‫for example, CodeCommit or GitHub,

16
00:00:38,497 --> 00:00:42,132
‫then the CodeGuru can have a look at all the lines of code

17
00:00:42,132 --> 00:00:44,900
‫and then can give you actionable recommendations

18
00:00:44,900 --> 00:00:47,480
‫in case it detects a bug or a memory leak

19
00:00:47,480 --> 00:00:49,310
‫or something that it has seen before.

20
00:00:49,310 --> 00:00:52,220
‫So, because of this machine learning capability,

21
00:00:52,220 --> 00:00:55,660
‫it can detect bugs before even other reviewers detect them,

22
00:00:55,660 --> 00:00:56,990
‫which is very helpful.

23
00:00:56,990 --> 00:00:59,257
‫And then CodeGuru Profiler is going to be here

24
00:00:59,257 --> 00:01:01,531
‫to give you visibility or recommendations

25
00:01:01,531 --> 00:01:03,600
‫about your application performance

26
00:01:03,600 --> 00:01:05,006
‫during runtimes or in production.

27
00:01:05,006 --> 00:01:07,205
‫So when you build and test your application,

28
00:01:07,205 --> 00:01:09,900
‫CodeGuru Profiler is already going to detect

29
00:01:09,900 --> 00:01:13,730
‫and optimize the expensive lines of code pre-production.

30
00:01:13,730 --> 00:01:15,570
‫And then when you deploy your application,

31
00:01:15,570 --> 00:01:17,980
‫you're going to measure your application in real time,

32
00:01:17,980 --> 00:01:20,210
‫and CodeGuru Profiler, yet again,

33
00:01:20,210 --> 00:01:21,800
‫is going to identify performance

34
00:01:21,800 --> 00:01:23,493
‫and cost improvements in production

35
00:01:23,493 --> 00:01:26,840
‫and give you these recommendations directly in your code.

36
00:01:26,840 --> 00:01:28,443
‫So this is the whole power CodeGuru.

37
00:01:28,443 --> 00:01:30,187
‫So if we do a deep dive,

38
00:01:30,187 --> 00:01:33,500
‫CodeGuru Reviewer really looks at your commits,

39
00:01:33,500 --> 00:01:35,470
‫so whenever you push your code,

40
00:01:35,470 --> 00:01:39,167
‫and tells you the lines of code that are probably wrong,

41
00:01:39,167 --> 00:01:40,990
‫so it could be very, very handy,

42
00:01:40,990 --> 00:01:42,940
‫so you can identify critical issues,

43
00:01:42,940 --> 00:01:45,180
‫security vulnerabilities, and hard-to-find bugs.

44
00:01:45,180 --> 00:01:48,491
‫So, for example, you can implement coding best practices,

45
00:01:48,491 --> 00:01:51,530
‫you can find resource leaks, do security detections

46
00:01:51,530 --> 00:01:53,910
‫in case you're creating a security hole,

47
00:01:53,910 --> 00:01:55,768
‫or input validation.

48
00:01:55,768 --> 00:01:58,510
‫The way it does it is using machine learning

49
00:01:58,510 --> 00:01:59,637
‫and automated reasoning.

50
00:01:59,637 --> 00:02:00,900
‫And how does it do it?

51
00:02:00,900 --> 00:02:04,410
‫Well, there are code reviews that were analyzed by CodeGuru

52
00:02:04,410 --> 00:02:07,460
‫across thousands of open source repositories out there,

53
00:02:07,460 --> 00:02:10,730
‫and also on all the amazon.com repositories,

54
00:02:10,730 --> 00:02:13,612
‫so this is how it learned to be a code reviewer

55
00:02:13,612 --> 00:02:15,060
‫through machine learning.

56
00:02:15,060 --> 00:02:16,900
‫It supports currently Java and Python,

57
00:02:16,900 --> 00:02:18,233
‫and it has integration with GitHub,

58
00:02:18,233 --> 00:02:19,981
‫Bitbucket, and CodeCommit.

59
00:02:19,981 --> 00:02:21,480
‫Now, don't quote me on this.

60
00:02:21,480 --> 00:02:22,774
‫Maybe things will evolve over time.

61
00:02:22,774 --> 00:02:24,370
‫If it does evolve, don't worry:

62
00:02:24,370 --> 00:02:26,046
‫you don't need to know that level of detail.

63
00:02:26,046 --> 00:02:28,566
‫Just know about CodeGuru, CodeGuru Reviewer,

64
00:02:28,566 --> 00:02:31,035
‫and then, lastly, CodeGuru Profiler.

65
00:02:31,035 --> 00:02:34,254
‫So Profiler is when your application is in production

66
00:02:34,254 --> 00:02:35,630
‫or in pre-prod,

67
00:02:35,630 --> 00:02:37,310
‫and it helps understand the runtime behavior

68
00:02:37,310 --> 00:02:38,470
‫of your application,

69
00:02:38,470 --> 00:02:42,148
‫and to look at what consumes excessive CPU capacity,

70
00:02:42,148 --> 00:02:44,650
‫for example, on the logging routine.

71
00:02:44,650 --> 00:02:46,565
‫So it will allow you to identify

72
00:02:46,565 --> 00:02:48,525
‫and remove code inefficiencies,

73
00:02:48,525 --> 00:02:50,220
‫improve the application performance,

74
00:02:50,220 --> 00:02:51,996
‫for example, reduce the CPU utilization,

75
00:02:51,996 --> 00:02:54,930
‫decrease compute costs, provide heap summaries

76
00:02:54,930 --> 00:02:56,413
‫so to identify which objects

77
00:02:56,413 --> 00:02:58,830
‫are taking a lot of space in memory,

78
00:02:58,830 --> 00:03:00,630
‫anomaly detection, as well,

79
00:03:00,630 --> 00:03:02,721
‫in case your application behaves weirdly.

80
00:03:02,721 --> 00:03:04,460
‫It also supports applications

81
00:03:04,460 --> 00:03:07,975
‫that will be running on AWS Cloud or even on-premises.

82
00:03:07,975 --> 00:03:09,120
‫And there's going to be

83
00:03:09,120 --> 00:03:10,720
‫a minimal overhead on the application

84
00:03:10,720 --> 00:03:13,150
‫to be monitored using CodeGuru Profiler.

85
00:03:13,150 --> 00:03:14,880
‫So that's it for this lecture.

86
00:03:14,880 --> 00:03:16,932
‫Just, again, remember the high level of CodeGuru,

87
00:03:16,932 --> 00:03:19,220
‫CodeGuru Reviewer, and Profiler,

88
00:03:19,220 --> 00:03:20,221
‫and you should be good to go.

89
00:03:20,221 --> 00:03:22,263
‫And I will see you in the next lecture.