1
00:00:06,920 --> 00:00:09,860
- In an AWS CI CD pipeline,

2
00:00:09,860 --> 00:00:12,710
there are a number of
services that you can use,

3
00:00:12,710 --> 00:00:17,710
either individually or all in
a single end to end process.

4
00:00:20,380 --> 00:00:23,720
We have code commit to
store your source code,

5
00:00:23,720 --> 00:00:25,610
there's code build,

6
00:00:25,610 --> 00:00:29,420
which builds deployment
artifacts and performs testing,

7
00:00:29,420 --> 00:00:31,950
we have code pipeline,

8
00:00:31,950 --> 00:00:35,400
which can orchestrate
the end to end process,

9
00:00:35,400 --> 00:00:38,445
and then we have this service CodeDeploy,

10
00:00:38,445 --> 00:00:41,830
which is designed for the last piece

11
00:00:41,830 --> 00:00:44,950
which is actually deploying
the code with a degree

12
00:00:44,950 --> 00:00:48,683
of quality and following
good DevOps principles.

13
00:00:49,576 --> 00:00:53,400
CodeDeploy, like the other
services is region scoped.

14
00:00:53,400 --> 00:00:56,420
And it's name is exactly what it does,

15
00:00:56,420 --> 00:01:00,493
it is a managed artifact
deployment service.

16
00:01:02,100 --> 00:01:06,840
It is designed around the
automation of software deployment

17
00:01:06,840 --> 00:01:10,610
to increase the uptime of the application

18
00:01:10,610 --> 00:01:15,333
and reduce the impact on
the end users or clients.

19
00:01:17,710 --> 00:01:20,580
CodeDeploy has native integration

20
00:01:20,580 --> 00:01:22,940
with the EC2 auto scaling.

21
00:01:22,940 --> 00:01:25,520
And so even if you have
an auto scaling group,

22
00:01:25,520 --> 00:01:27,660
with a number of running EC2 instances

23
00:01:27,660 --> 00:01:29,100
and you need to update the version

24
00:01:29,100 --> 00:01:30,520
of the application on them,

25
00:01:30,520 --> 00:01:32,670
you can do that using CodeDeploy.

26
00:01:34,220 --> 00:01:36,120
It has on-prem capability,

27
00:01:36,120 --> 00:01:39,980
so you can perform these
deployments against resources

28
00:01:39,980 --> 00:01:42,130
that are sitting in an
on-prem data center.

29
00:01:43,780 --> 00:01:46,910
It allows for different
types of deployments,

30
00:01:46,910 --> 00:01:51,613
including a rolling deployment
and blue/green deployments.

31
00:01:53,370 --> 00:01:57,780
It also allows for configuration
of deployment groups

32
00:01:57,780 --> 00:01:59,880
as well as handling the history

33
00:01:59,880 --> 00:02:03,683
of deployments as well as
success and failure rates.

34
00:02:05,850 --> 00:02:09,330
And finally it integrates
with the other services

35
00:02:09,330 --> 00:02:12,347
that can be part of a
CI CD pipeline in AWS.

36
00:02:15,390 --> 00:02:17,560
Let's take a look at an example

37
00:02:17,560 --> 00:02:21,560
of using CodeDeploy all by itself

38
00:02:21,560 --> 00:02:26,560
for managing the application
version of a Lambda function.

39
00:02:27,890 --> 00:02:29,310
So, first thing we need to do

40
00:02:29,310 --> 00:02:33,100
is we create the function
itself, the application,

41
00:02:33,100 --> 00:02:36,050
and we are going to associate that

42
00:02:36,050 --> 00:02:41,050
with a deployment group,
configuration and versions.

43
00:02:42,100 --> 00:02:45,040
And this will help us to maintain

44
00:02:45,040 --> 00:02:49,740
state within CodeDeploy.

45
00:02:49,740 --> 00:02:52,470
Next, we specify the deployment group

46
00:02:52,470 --> 00:02:54,190
that we want to deploy to

47
00:02:54,190 --> 00:02:57,020
in this particular iteration.

48
00:02:57,020 --> 00:02:59,110
And this is where we can
configure things like

49
00:02:59,110 --> 00:03:02,910
traffic routing between
multiple deployment groups,

50
00:03:02,910 --> 00:03:06,140
so that we can perform gradual cut overs

51
00:03:06,140 --> 00:03:09,710
from one version of an
application to another,

52
00:03:09,710 --> 00:03:12,273
and it allows us to support rollback.

53
00:03:14,730 --> 00:03:17,320
Next, we have the
deployment configuration,

54
00:03:17,320 --> 00:03:20,150
this is where we choose
the actual algorithm

55
00:03:20,150 --> 00:03:21,020
for the deployment.

56
00:03:21,020 --> 00:03:22,860
Do we do it all at once?

57
00:03:22,860 --> 00:03:25,010
Do we do a Canary deployment?

58
00:03:25,010 --> 00:03:27,403
Or do we do a blue/green
deployment instead?

59
00:03:30,120 --> 00:03:34,450
The AppSpec file is where
we actually describe

60
00:03:34,450 --> 00:03:36,940
the function versions, the current version

61
00:03:36,940 --> 00:03:40,310
and the version that we want to deploy to

62
00:03:40,310 --> 00:03:43,423
along with other validation operations.

63
00:03:45,980 --> 00:03:48,770
Now we reach the actual deploy step

64
00:03:48,770 --> 00:03:53,700
and this allows us to push
that new application revision

65
00:03:53,700 --> 00:03:56,070
to the deployment group using

66
00:03:56,070 --> 00:03:57,510
the method that we've defined

67
00:03:57,510 --> 00:03:59,120
in the deployment configuration,

68
00:03:59,120 --> 00:04:01,490
which could be just a Canary deployment,

69
00:04:01,490 --> 00:04:04,940
meaning we have a very small percentage

70
00:04:04,940 --> 00:04:09,940
of our actual application
that is using that new version

71
00:04:10,010 --> 00:04:12,403
of the function until
we're sure that it works.

72
00:04:14,586 --> 00:04:16,460
We then check our results

73
00:04:16,460 --> 00:04:18,340
and this is going to use CloudWatch

74
00:04:18,340 --> 00:04:20,853
and other built in AWS monitoring.

75
00:04:21,740 --> 00:04:24,260
And finally, we make the decision,

76
00:04:24,260 --> 00:04:25,730
this is our tollgate,

77
00:04:25,730 --> 00:04:30,090
where we decide was the
deployment truly successful

78
00:04:30,090 --> 00:04:33,440
or did we need to trigger a rollback so

79
00:04:33,440 --> 00:04:35,730
that we can go back to
the previous version

80
00:04:35,730 --> 00:04:40,163
without any further impact
on our end users or clients.