1
00:00:00,090 --> 00:00:01,380
‫So before we go ahead

2
00:00:01,380 --> 00:00:03,450
‫and practice creating an auto scaling group,

3
00:00:03,450 --> 00:00:05,790
‫you need to take your first two instances

4
00:00:05,790 --> 00:00:07,940
‫and we're actually going to terminate them.

5
00:00:08,820 --> 00:00:10,110
‫Okay, so now this is done,

6
00:00:10,110 --> 00:00:12,900
‫we can go ahead and create an auto scaling group.

7
00:00:12,900 --> 00:00:16,020
‫For this, on the bottom left, click on Auto Scaling Group

8
00:00:16,020 --> 00:00:18,450
‫and we will create an auto scaling group.

9
00:00:18,450 --> 00:00:21,090
‫So I'll call this one DemoASG,

10
00:00:21,090 --> 00:00:23,730
‫and we need to create a launch template.

11
00:00:23,730 --> 00:00:24,660
‫So currently we have none.

12
00:00:24,660 --> 00:00:27,033
‫So let's create a launch template,

13
00:00:29,340 --> 00:00:32,700
‫and I will call this one DemoLaunchTemplate.

14
00:00:32,700 --> 00:00:35,490
‫And this template is being used to tell

15
00:00:35,490 --> 00:00:39,300
‫to the ASG how to create EC2 instances within it.

16
00:00:39,300 --> 00:00:41,250
‫So this will look very, very similar

17
00:00:41,250 --> 00:00:44,010
‫to what we have when we create EC2 instances.

18
00:00:44,010 --> 00:00:47,010
‫As you can see here, I can choose, for example,

19
00:00:47,010 --> 00:00:50,940
‫a quick start Amazon Linux for getting Amazon Linux 2

20
00:00:50,940 --> 00:00:53,640
‫as the base of my EC2 instance.

21
00:00:53,640 --> 00:00:56,160
‫Then we have an instant type that we can include,

22
00:00:56,160 --> 00:00:58,620
‫for example, t2.micro.

23
00:00:58,620 --> 00:01:00,900
‫For key pair, we will not include it in the launch templates

24
00:01:00,900 --> 00:01:03,480
‫or we can just say that, no, we don't need one,

25
00:01:03,480 --> 00:01:04,950
‫so this is good enough.

26
00:01:04,950 --> 00:01:06,570
‫For subnets, so we'll not include this

27
00:01:06,570 --> 00:01:07,740
‫in the launch templates.

28
00:01:07,740 --> 00:01:09,840
‫For security group, we can select

29
00:01:09,840 --> 00:01:12,030
‫a security group that's already existing.

30
00:01:12,030 --> 00:01:13,863
‫For example, my launch-wizard-1,

31
00:01:16,620 --> 00:01:18,390
‫under advanced network configuration

32
00:01:18,390 --> 00:01:19,650
‫we don't need to do anything.

33
00:01:19,650 --> 00:01:22,740
‫For EBS volumes for storage we don't need to do anything.

34
00:01:22,740 --> 00:01:24,720
‫And then for advanced details

35
00:01:24,720 --> 00:01:27,540
‫we want these instances to start with some user data.

36
00:01:27,540 --> 00:01:29,670
‫And so we scroll all the way down

37
00:01:29,670 --> 00:01:34,110
‫and here we copy and paste the user data.

38
00:01:34,110 --> 00:01:36,360
‫Okay, so let's create this launch template.

39
00:01:36,360 --> 00:01:38,490
‫As you can see, thanks to this launch template

40
00:01:38,490 --> 00:01:41,310
‫we launch EC2 instances just like before.

41
00:01:41,310 --> 00:01:42,630
‫So let's refresh this

42
00:01:42,630 --> 00:01:46,110
‫and then click on the DemoLaunchTemplate of version 1.

43
00:01:46,110 --> 00:01:49,380
‫So here it describes what is going to happen, the type

44
00:01:49,380 --> 00:01:51,900
‫of instance we're going to have, the security groups

45
00:01:51,900 --> 00:01:53,190
‫and so on.

46
00:01:53,190 --> 00:01:54,570
‫So let's click on next.

47
00:01:54,570 --> 00:01:57,840
‫Next we need to choose where to launch these instances.

48
00:01:57,840 --> 00:01:59,280
‫So we have our VPC,

49
00:01:59,280 --> 00:02:01,290
‫and then we can select multiple

50
00:02:01,290 --> 00:02:02,700
‫availability zones and subnets.

51
00:02:02,700 --> 00:02:04,950
‫So we select three of them.

52
00:02:04,950 --> 00:02:08,010
‫And for instance type requirements, we can use the one

53
00:02:08,010 --> 00:02:10,680
‫from the launch template or if you wanted to override them

54
00:02:10,680 --> 00:02:11,970
‫but we don't need to actually.

55
00:02:11,970 --> 00:02:13,120
‫So let's click on next.

56
00:02:14,550 --> 00:02:15,840
‫Next we have load balancers.

57
00:02:15,840 --> 00:02:19,090
‫So we want to attach to an existing load balancer

58
00:02:19,980 --> 00:02:22,650
‫and it's going to be a application balancer.

59
00:02:22,650 --> 00:02:26,130
‫And to do this, we're going to tell the ASG,

60
00:02:26,130 --> 00:02:29,640
‫the auto scanning group, that every instance created should

61
00:02:29,640 --> 00:02:34,410
‫be registered within my demo target group for my ALB.

62
00:02:34,410 --> 00:02:36,360
‫So therefore all these instances will be

63
00:02:36,360 --> 00:02:37,650
‫under the target group,

64
00:02:37,650 --> 00:02:39,060
‫and then the load balancer will be able

65
00:02:39,060 --> 00:02:41,220
‫to direct traffic to them.

66
00:02:41,220 --> 00:02:44,670
‫So the health check can be EC2 and also ELB.

67
00:02:44,670 --> 00:02:45,870
‫So we're good.

68
00:02:45,870 --> 00:02:48,453
‫And then we can click on next.

69
00:02:49,710 --> 00:02:53,190
‫Now here comes the scaling of the auto scaling group.

70
00:02:53,190 --> 00:02:56,190
‫So the desired capacity is how many instances you want

71
00:02:56,190 --> 00:02:57,023
‫at any time.

72
00:02:57,023 --> 00:02:58,620
‫For example, we want two EC2 instances

73
00:02:58,620 --> 00:03:00,780
‫to have some sort of load balancing.

74
00:03:00,780 --> 00:03:02,100
‫The minimum capacity is one,

75
00:03:02,100 --> 00:03:04,680
‫meaning we want at least one instance,

76
00:03:04,680 --> 00:03:08,490
‫and the maximum for them is we want at most four instances.

77
00:03:08,490 --> 00:03:10,050
‫But the desired is what matters the most

78
00:03:10,050 --> 00:03:13,560
‫because this is the actual capacity that we're going to get.

79
00:03:13,560 --> 00:03:16,170
‫Okay, so then do we want scaling policies?

80
00:03:16,170 --> 00:03:18,510
‫This is too advanced, but you of course can set

81
00:03:18,510 --> 00:03:21,090
‫scaling policies on a auto scaling group.

82
00:03:21,090 --> 00:03:22,470
‫That's the whole point of it

83
00:03:22,470 --> 00:03:27,420
‫which allows you to resize your ASG on demand.

84
00:03:27,420 --> 00:03:30,180
‫If there is much more demand than is going to

85
00:03:30,180 --> 00:03:33,570
‫have more instances, if there is less requests,

86
00:03:33,570 --> 00:03:37,113
‫less demand than is going to be less instances.

87
00:03:38,160 --> 00:03:39,630
‫So click on next.

88
00:03:39,630 --> 00:03:42,240
‫Next, we don't need notifications, we don't need tags.

89
00:03:42,240 --> 00:03:43,500
‫We can review everything.

90
00:03:43,500 --> 00:03:44,333
‫It looks good.

91
00:03:44,333 --> 00:03:46,443
‫And let's create our auto scaling group.

92
00:03:47,820 --> 00:03:51,720
‫So now our auto scaling group is being created

93
00:03:51,720 --> 00:03:54,630
‫and as you can see, the state is updating capacity

94
00:03:54,630 --> 00:03:58,650
‫because we have zero instances in our ASG, but we want to,

95
00:03:58,650 --> 00:04:01,650
‫so I can click on it to get a bit more details.

96
00:04:01,650 --> 00:04:03,783
‫So let's go under activity.

97
00:04:04,710 --> 00:04:07,500
‫And in here we have two activity history.

98
00:04:07,500 --> 00:04:11,490
‫That is we are launching two new EC2 instances

99
00:04:11,490 --> 00:04:15,450
‫because well, the desired capacity went from zero to two.

100
00:04:15,450 --> 00:04:16,950
‫And so if we have a look

101
00:04:16,950 --> 00:04:20,670
‫under the instance management tab, as you can see now,

102
00:04:20,670 --> 00:04:24,300
‫2 EC2 instances are independent states.

103
00:04:24,300 --> 00:04:28,380
‫So if I go under EC2 and look at my EC2 instances,

104
00:04:28,380 --> 00:04:31,980
‫in that UI, we also see that two instances are running

105
00:04:31,980 --> 00:04:35,640
‫and these have been created by my auto scaling group.

106
00:04:35,640 --> 00:04:38,370
‫So the benefit is that now they are fully managed

107
00:04:38,370 --> 00:04:40,170
‫by my auto scaling group,

108
00:04:40,170 --> 00:04:43,470
‫and let's go see for example, in my target group as well.

109
00:04:43,470 --> 00:04:46,230
‫So if I go to my target group on the left hand side

110
00:04:46,230 --> 00:04:50,190
‫and look at my demo-tg-alb right here, scroll down.

111
00:04:50,190 --> 00:04:52,740
‫As you can see now we have two total targets

112
00:04:52,740 --> 00:04:55,380
‫and these are the targets created

113
00:04:55,380 --> 00:04:57,210
‫by our auto scanning group.

114
00:04:57,210 --> 00:05:00,210
‫So again, thanks to the integration that we've defined

115
00:05:00,210 --> 00:05:01,650
‫between the auto scanning group

116
00:05:01,650 --> 00:05:06,090
‫and the load balancer, we are able to have automatically

117
00:05:06,090 --> 00:05:09,360
‫these new EC2 instances registered as targets

118
00:05:09,360 --> 00:05:11,370
‫in our target group.

119
00:05:11,370 --> 00:05:13,260
‫So currently they're unhealthy,

120
00:05:13,260 --> 00:05:16,020
‫this is because the instance hasn't started all the way yet

121
00:05:16,020 --> 00:05:18,670
‫so let's wait a little bit until they become healthy.

122
00:05:20,580 --> 00:05:23,580
‫So to speed up the check from unhealthy to healthy,

123
00:05:23,580 --> 00:05:26,640
‫you can go under health checks of your target group,

124
00:05:26,640 --> 00:05:29,070
‫and here you can edit these settings.

125
00:05:29,070 --> 00:05:30,690
‫And under other settings, we can say

126
00:05:30,690 --> 00:05:35,040
‫that the healthy threshold is going to be 2

127
00:05:35,040 --> 00:05:38,670
‫and the interval is going to be 5 seconds.

128
00:05:38,670 --> 00:05:41,910
‫This is going to make the thing much quicker.

129
00:05:41,910 --> 00:05:44,160
‫So of course the timeout needs to be 2 seconds,

130
00:05:44,160 --> 00:05:46,860
‫something less than the interval itself.

131
00:05:46,860 --> 00:05:48,333
‫So let's save our changes.

132
00:05:49,620 --> 00:05:52,020
‫And now the health check change settings have changed.

133
00:05:52,020 --> 00:05:54,060
‫Let's see, if I go back into my targets

134
00:05:54,060 --> 00:05:57,000
‫and refresh, now both my instances are healthy.

135
00:05:57,000 --> 00:06:00,900
‫We just made the health check happen faster and more often.

136
00:06:00,900 --> 00:06:03,120
‫So now both my instances are healthy,

137
00:06:03,120 --> 00:06:08,120
‫and so therefore if I go under my load balancer right here

138
00:06:08,280 --> 00:06:10,020
‫and I look at the DNS name

139
00:06:10,020 --> 00:06:12,930
‫and open it in a new tab, I get a hello world

140
00:06:12,930 --> 00:06:14,910
‫from both my instances.

141
00:06:14,910 --> 00:06:18,540
‫And this is cool because these two instances were created

142
00:06:18,540 --> 00:06:20,490
‫by the auto scaling group.

143
00:06:20,490 --> 00:06:22,590
‫So because now we have an auto scaling group

144
00:06:22,590 --> 00:06:24,450
‫we can actually do some cool stuff.

145
00:06:24,450 --> 00:06:27,480
‫So if we take one of these instances, for example

146
00:06:27,480 --> 00:06:29,280
‫and we can for example terminate it,

147
00:06:29,280 --> 00:06:30,790
‫so I'm going to click on it

148
00:06:33,810 --> 00:06:35,880
‫and I'm under the instance itself

149
00:06:35,880 --> 00:06:38,853
‫I will do instance date and then terminate instance.

150
00:06:41,670 --> 00:06:43,800
‫Now it's been successfully terminated.

151
00:06:43,800 --> 00:06:45,030
‫So what's going to happen is that

152
00:06:45,030 --> 00:06:47,850
‫because this instance is being shutting down

153
00:06:47,850 --> 00:06:50,460
‫and terminated, well, my auto scaling group

154
00:06:50,460 --> 00:06:53,520
‫is going to detect that, guess what, one of these instance

155
00:06:53,520 --> 00:06:56,310
‫is not in service anymore, it's being terminated.

156
00:06:56,310 --> 00:06:59,070
‫And so therefore, because we have an auto scaling group

157
00:06:59,070 --> 00:07:02,730
‫with a desired capacity of two instances,

158
00:07:02,730 --> 00:07:05,670
‫automatically a new instance should appear.

159
00:07:05,670 --> 00:07:07,680
‫So let's observe this behavior

160
00:07:07,680 --> 00:07:11,610
‫by having here the activity history.

161
00:07:11,610 --> 00:07:12,600
‫And as you can see,

162
00:07:12,600 --> 00:07:15,870
‫in progress was terminating EC2 instance.

163
00:07:15,870 --> 00:07:17,970
‫And so an instance was taken out of service

164
00:07:17,970 --> 00:07:20,730
‫because, well, it's been terminated.

165
00:07:20,730 --> 00:07:23,430
‫And then we have a new activity saying,

166
00:07:23,430 --> 00:07:26,340
‫hey, an instance was launched in response

167
00:07:26,340 --> 00:07:28,830
‫to an unhealthy instance needing to be replaced.

168
00:07:28,830 --> 00:07:30,900
‫So it's very cool because the auto scaling group

169
00:07:30,900 --> 00:07:33,720
‫can automatically detect unhealthy instances

170
00:07:33,720 --> 00:07:36,390
‫and create new one for replacements.

171
00:07:36,390 --> 00:07:39,330
‫So if we go here now, there's one instance in pending state

172
00:07:39,330 --> 00:07:42,660
‫which is being started, one instance being terminated

173
00:07:42,660 --> 00:07:44,610
‫and one instance in service.

174
00:07:44,610 --> 00:07:47,790
‫And this is the whole power of auto scaling groups.

175
00:07:47,790 --> 00:07:49,290
‫Of course you can go to the next level

176
00:07:49,290 --> 00:07:50,550
‫but for now we know enough

177
00:07:50,550 --> 00:07:52,620
‫which is around automatic scaling to actually

178
00:07:52,620 --> 00:07:55,680
‫define scaling policies to automatically increase

179
00:07:55,680 --> 00:07:57,960
‫or decrease the desired capacity over time,

180
00:07:57,960 --> 00:08:00,120
‫based on our load and so on.

181
00:08:00,120 --> 00:08:02,580
‫But here you've seen the basics

182
00:08:02,580 --> 00:08:05,310
‫and the major features of auto scaling groups,

183
00:08:05,310 --> 00:08:06,810
‫and you could play around

184
00:08:06,810 --> 00:08:09,570
‫by editing the desired capacity yourself

185
00:08:09,570 --> 00:08:12,000
‫to set it to one, for example, to terminate instances

186
00:08:12,000 --> 00:08:14,790
‫and only keep one of them, or to set it to four

187
00:08:14,790 --> 00:08:17,820
‫and see the auto scaling group creates multiple instances

188
00:08:17,820 --> 00:08:20,670
‫that will be registered with our load balancer.

189
00:08:20,670 --> 00:08:23,400
‫And so therefore the traffic is going to be distributed

190
00:08:23,400 --> 00:08:24,980
‫between four instances.

191
00:08:24,980 --> 00:08:26,160
‫So I hope you liked it

192
00:08:26,160 --> 00:08:28,110
‫and I will see you in the next lecture.