1
00:00:02,890 --> 00:00:05,340
‫There are two pieces of metadata in this YAML that

2
00:00:06,160 --> 00:00:08,218
‫you probably have seen in the examples and

3
00:00:09,250 --> 00:00:12,249
‫thought about, wondering what labels and annotations

4
00:00:12,250 --> 00:00:13,250
‫are really there for.

5
00:00:13,930 --> 00:00:15,910
‫You probably didn't realize their significance.

6
00:00:15,940 --> 00:00:18,999
‫They're a key component of Kubernetes and

7
00:00:19,000 --> 00:00:20,437
‫how things connect to each other.

8
00:00:20,438 --> 00:00:21,879
‫Much more so than Swarm.

9
00:00:22,580 --> 00:00:25,600
‫The first one is labels. Labels go under the metadata

10
00:00:25,930 --> 00:00:28,980
‫key in the root of the file, so they're

11
00:00:28,990 --> 00:00:31,685
‫optional there. You don't have to put them but they are

12
00:00:32,200 --> 00:00:35,439
‫a key value, or a list of key values, underneath

13
00:00:35,560 --> 00:00:36,640
‫the labels heading.

14
00:00:36,850 --> 00:00:38,950
‫Those are there for multiple reasons.

15
00:00:39,190 --> 00:00:42,460
‫They're not always required, but you may need them for

16
00:00:42,520 --> 00:00:45,699
‫selecting the objects, grouping them

17
00:00:45,730 --> 00:00:48,129
‫and even filtering them later on.

18
00:00:48,340 --> 00:00:50,251
‫This can be anything that describes the

19
00:00:51,310 --> 00:00:52,569
‫object or the resource.

20
00:00:52,600 --> 00:00:55,899
‫It can be a key value pair that the value

21
00:00:55,900 --> 00:00:57,039
‫is simply true or false.

22
00:00:57,340 --> 00:00:58,869
‫The keys can be whatever you want.

23
00:00:58,900 --> 00:01:01,960
‫There just can't be more than one key in each object,

24
00:01:01,990 --> 00:01:04,869
‫so you can't have two keys in a row in the YAML

25
00:01:05,200 --> 00:01:07,869
‫for the same object that have the key name of, let's say

26
00:01:07,870 --> 00:01:11,499
‫app, right. These are used really to describe

27
00:01:11,710 --> 00:01:14,500
‫the object. These resources that we're creating

28
00:01:15,160 --> 00:01:16,571
‫have different properties to them.

29
00:01:16,572 --> 00:01:19,069
‫Maybe it's the environment like prod, dev or test.

30
00:01:19,180 --> 00:01:21,760
‫Maybe it's the location like dmz, or

31
00:01:22,130 --> 00:01:25,239
‫backend or frontend. Maybe it's the app name, or

32
00:01:25,240 --> 00:01:28,120
‫maybe it's the client name, if you're hosting client sites

33
00:01:28,150 --> 00:01:30,339
‫or something. It could just be whatever you want and

34
00:01:30,340 --> 00:01:31,840
‫there's not a whole lot of standards.

35
00:01:32,170 --> 00:01:34,800
‫The Kubernetes documentation listed in the resources here

36
00:01:34,820 --> 00:01:37,440
‫does give you some examples, but those are just examples.

37
00:01:37,450 --> 00:01:39,849
‫There's not a whole lot of standards in the industry.

38
00:01:39,850 --> 00:01:42,699
‫It's really up to your team, and your company, about how

39
00:01:42,700 --> 00:01:44,919
‫you want to describe this thing.

40
00:01:45,290 --> 00:01:48,400
‫We also have annotations. They're sort of a yin and yang.

41
00:01:48,430 --> 00:01:50,170
‫The labels are very simple.

42
00:01:50,230 --> 00:01:52,929
‫They are very limited in size and what characters you can

43
00:01:52,930 --> 00:01:55,930
‫put in them. They're meant to describe the resource.

44
00:01:56,320 --> 00:01:59,890
‫With annotations, this is for more complex

45
00:02:00,070 --> 00:02:02,919
‫data and multi-value data, stuff like that.

46
00:02:03,160 --> 00:02:06,099
‫You'll usually see annotations used when

47
00:02:06,130 --> 00:02:09,039
‫you're adding custom stuff into Kubernetes that

48
00:02:09,160 --> 00:02:11,560
‫can talk to Kubernetes and get data out of it.

49
00:02:11,800 --> 00:02:14,650
‫The annotations are there to maybe store config data, like

50
00:02:14,860 --> 00:02:17,919
‫proxies and ingress stuff can often store

51
00:02:17,920 --> 00:02:20,949
‫different configurations, inside different annotations.

52
00:02:21,220 --> 00:02:23,919
‫So, those are not necessarily there

53
00:02:24,190 --> 00:02:27,550
‫for describing but more for configuration.

54
00:02:27,820 --> 00:02:30,429
‫Let's go back to labels for a minute, because there's a lot

55
00:02:30,430 --> 00:02:33,321
‫more to that that's key to understanding how all this stuff

56
00:02:33,430 --> 00:02:34,430
‫works in Kubernetes.

57
00:02:34,870 --> 00:02:37,989
‫The first thing is that we can do simple filters

58
00:02:38,110 --> 00:02:40,509
‫by limiting to the labels from the command line.

59
00:02:40,540 --> 00:02:42,843
‫We can do a get pods and just say, only show me

60
00:02:43,550 --> 00:02:44,949
‫the pods with this label.

61
00:02:45,310 --> 00:02:47,919
‫You can even use a comma there to add multiple labels.

62
00:02:47,950 --> 00:02:51,039
‫In fact, there's a whole level of syntax

63
00:02:51,040 --> 00:02:54,015
‫to this thing where you can do different types of syntaxes,

64
00:02:54,033 --> 00:02:55,033
‫different or's and and's.

65
00:02:55,690 --> 00:02:57,699
‫Really, it's one of those things where I don't want to

66
00:02:57,700 --> 00:03:00,493
‫teach you all that because it's going to be like learning

67
00:03:00,610 --> 00:03:03,354
‫Regex. It's something where you won't really learn it by

68
00:03:03,550 --> 00:03:04,810
‫just trying it once.

69
00:03:04,840 --> 00:03:07,140
‫It's only when you need it that you're really going to

70
00:03:07,240 --> 00:03:09,699
‫understand it. The documentation from Kubernetes is really

71
00:03:09,700 --> 00:03:12,009
‫great. They've got multiple links I've put in the resources

72
00:03:12,370 --> 00:03:15,069
‫that will take you through the path of all the different

73
00:03:15,070 --> 00:03:16,719
‫ways that you might need labels.

74
00:03:16,990 --> 00:03:19,734
‫Quite frankly, you don't really need to learn that until

75
00:03:20,080 --> 00:03:22,479
‫you know you need it to solve a problem.

76
00:03:23,020 --> 00:03:24,849
‫You can even use it for apply commands.

77
00:03:24,910 --> 00:03:27,660
‫If you had a whole bunch of stuff in one big YAML

78
00:03:27,880 --> 00:03:30,183
‫file and you only wanted to apply certain parts

79
00:03:31,150 --> 00:03:32,150
‫of that YAML file.

80
00:03:32,720 --> 00:03:35,139
‫I don't do this a whole lot, but it is there where you

81
00:03:35,140 --> 00:03:38,380
‫could limit it to say, only apply

82
00:03:38,560 --> 00:03:41,360
‫the changes or information that you've put in the YAML

83
00:03:41,560 --> 00:03:42,560
‫for specific labels.

84
00:03:43,000 --> 00:03:44,859
‫Of course, again with this one, you can put the same

85
00:03:44,860 --> 00:03:46,869
‫syntax. You can just have a single label.

86
00:03:46,930 --> 00:03:49,509
‫You can put in commas, have multiple labels and values

87
00:03:49,510 --> 00:03:51,130
‫there. It's up to you.

88
00:03:51,640 --> 00:03:54,699
‫There's another thing that the labels do that's

89
00:03:54,790 --> 00:03:56,650
‫important to how Kubernetes works.

90
00:03:56,800 --> 00:03:59,979
‫It comes down to there are certain

91
00:04:00,280 --> 00:04:02,730
‫resources that talk to other resources, or connect

92
00:04:03,550 --> 00:04:05,229
‫to them, or are related to them.

93
00:04:05,440 --> 00:04:08,199
‫However you want to say it. The good example is the

94
00:04:08,200 --> 00:04:11,169
‫services need to know which pods to send their

95
00:04:11,170 --> 00:04:14,042
‫traffic to. You might wonder how does it know?

96
00:04:14,200 --> 00:04:17,199
‫I mean, when you type it in, if you do a kubectl service

97
00:04:17,200 --> 00:04:19,539
‫Command and you're pointing it to the deployment,

98
00:04:20,170 --> 00:04:22,767
‫what exactly is it using to identify that deployment?

99
00:04:22,830 --> 00:04:25,378
‫Well, in the CLI, you would think that it's actually

100
00:04:25,660 --> 00:04:27,069
‫pointing to the deployment object.

101
00:04:27,100 --> 00:04:29,746
‫But really, the service needs to talk to the pods, not

102
00:04:30,070 --> 00:04:31,070
‫the deployment itself.

103
00:04:31,660 --> 00:04:32,990
‫It's using something called label selectors.

104
00:04:32,991 --> 00:04:36,310
‫This is common with lots of the different resources

105
00:04:36,640 --> 00:04:39,129
‫that are higher levels of abstraction of pods.

106
00:04:39,730 --> 00:04:42,129
‫For them to relate to the pods, or to know which pods are

107
00:04:42,130 --> 00:04:44,860
‫their's, they use label selectors. You will see that

108
00:04:45,310 --> 00:04:47,760
‫in the YAML inside our repo that you've been using

109
00:04:48,370 --> 00:04:49,959
‫for the last few lectures.

110
00:04:50,320 --> 00:04:53,679
‫If you look in the app.yml, you'll notice that all along,

111
00:04:53,710 --> 00:04:55,899
‫we've had a couple of things in there that we haven't

112
00:04:55,900 --> 00:04:58,449
‫talked about. The first one is at the top part with the

113
00:04:58,450 --> 00:05:01,870
‫service, we have under the spec a selector.

114
00:05:02,500 --> 00:05:03,640
‫That has to be there.

115
00:05:04,000 --> 00:05:07,029
‫We're telling it the key and value of a label

116
00:05:07,330 --> 00:05:09,970
‫for pods that it needs to direct traffic to.

117
00:05:10,510 --> 00:05:12,009
‫That's just a simple match.

118
00:05:12,040 --> 00:05:13,779
‫It needs to find pods with that.

119
00:05:14,230 --> 00:05:17,259
‫Then down in the deployment underneath, you'll see that our

120
00:05:17,260 --> 00:05:19,930
‫spec has its own selector,

121
00:05:20,290 --> 00:05:22,750
‫matchLabels, app, app-nginx.

122
00:05:24,010 --> 00:05:27,160
‫What's happening there is both of those objects are

123
00:05:27,220 --> 00:05:29,980
‫selecting pods based on that match.

124
00:05:30,370 --> 00:05:33,309
‫That means in the template under our spec, where we're

125
00:05:33,310 --> 00:05:36,519
‫templating out the pods, that needs to have

126
00:05:36,580 --> 00:05:39,729
‫that label in it so that those resources

127
00:05:39,730 --> 00:05:40,779
‫can find their pods.

128
00:05:41,140 --> 00:05:44,049
‫The nice thing is in the later versions of Kubernetes, is

129
00:05:44,050 --> 00:05:47,139
‫that at least with just a single deployment, these have

130
00:05:47,140 --> 00:05:50,199
‫to be in sync. It'll actually fail to deploy if it

131
00:05:50,200 --> 00:05:53,589
‫sees that the selector match labels

132
00:05:53,680 --> 00:05:56,889
‫isn't matching the template metadata labels

133
00:05:57,100 --> 00:05:59,679
‫because they have to in order for it to even deploy the

134
00:05:59,680 --> 00:06:01,059
‫pods and be able to control them.

135
00:06:01,330 --> 00:06:03,969
‫It's there to kind of prevent you from accidentally doing

136
00:06:03,970 --> 00:06:06,249
‫it the wrong way. It will expect you to have those pieces

137
00:06:06,250 --> 00:06:08,740
‫in there. In older versions, there used to be a way where

138
00:06:08,741 --> 00:06:11,019
‫it would automatically generate them, by default, so you

139
00:06:11,020 --> 00:06:13,129
‫could leave the stuff out. They've stopped doing that now,

140
00:06:13,130 --> 00:06:15,489
‫and they force you to put this in there. I think that's a

141
00:06:15,490 --> 00:06:17,859
‫good thing, especially how important these labels selectors

142
00:06:17,860 --> 00:06:20,774
‫are. Technically, if you went back to Swarm, Swarm

143
00:06:21,460 --> 00:06:24,057
‫does a little bit of this out-of-the-box where it has

144
00:06:24,190 --> 00:06:26,738
‫labels that match things up from stacks to services,

145
00:06:27,730 --> 00:06:30,100
‫and stacks to their volumes, and stuff like that.

146
00:06:30,370 --> 00:06:32,379
‫We don't really get to control it, and we don't see it a

147
00:06:32,380 --> 00:06:33,909
‫whole lot in the interface. It's kind of done in the

148
00:06:33,910 --> 00:06:34,910
‫background for us in Swarm.

149
00:06:35,500 --> 00:06:37,120
‫But, it's also very limiting.

150
00:06:37,150 --> 00:06:39,629
‫You can't change things on the fly.

151
00:06:39,640 --> 00:06:42,579
‫You can't necessarily go and customize those for really

152
00:06:42,580 --> 00:06:43,580
‫unique setups.

153
00:06:43,890 --> 00:06:46,959
‫Here, we could possibly have a

154
00:06:47,080 --> 00:06:49,138
‫service talk to multiple deployments if we

155
00:06:50,110 --> 00:06:51,610
‫wanted to change those selectors.

156
00:06:52,000 --> 00:06:54,819
‫Selectors aren't just about linking resources together.

157
00:06:54,880 --> 00:06:57,069
‫It's also about controlling where they go.

158
00:06:57,190 --> 00:07:00,309
‫On the nodes, you can use labels and selectors

159
00:07:00,310 --> 00:07:01,660
‫to control where they go.

160
00:07:02,050 --> 00:07:05,080
‫You can use even further customization with things called

161
00:07:05,110 --> 00:07:07,560
‫taints and toleration that are sort of the inverse

162
00:07:08,320 --> 00:07:11,499
‫of that where you're telling nodes that they can't

163
00:07:11,500 --> 00:07:13,852
‫do certain things so that Kubernetes knows where

164
00:07:14,470 --> 00:07:16,449
‫to basically control where your pods go.

165
00:07:16,900 --> 00:07:19,899
‫These things get really custom and really advanced

166
00:07:19,990 --> 00:07:23,260
‫pretty quickly. Doing them all together can get

167
00:07:23,370 --> 00:07:25,059
‫a little bit hairy and crazy.

168
00:07:25,330 --> 00:07:28,330
‫I definitely recommend that you take it slow with them.

169
00:07:28,420 --> 00:07:29,740
‫Keep it all simple out-of-the-box.

170
00:07:30,430 --> 00:07:33,459
‫Avoid taints and tolerations until you really

171
00:07:33,460 --> 00:07:34,629
‫need to look into those.

172
00:07:35,140 --> 00:07:37,329
‫All right. Let's clean up for this whole section.

173
00:07:37,360 --> 00:07:40,479
‫If you have anything running in a default namespace

174
00:07:40,480 --> 00:07:43,371
‫for your cluster, if you just do a kubectl get all, and you

175
00:07:43,510 --> 00:07:46,552
‫see stuff there beyond the Kubernetes API itself, then

176
00:07:47,380 --> 00:07:50,429
‫go ahead and delete all those resources because we're

177
00:07:50,440 --> 00:07:51,760
‫going to be moving on. We won't need those.