1 00:00:02,280 --> 00:00:05,640 So now the cluster and the service we started 2 00:00:05,640 --> 00:00:09,660 in ECS was deleted, now we can create a new cluster, 3 00:00:09,660 --> 00:00:14,660 new tasks, new services to run our backend container here. 4 00:00:15,240 --> 00:00:20,240 So for that I'll use this UI, click on Create Cluster 5 00:00:20,750 --> 00:00:24,513 and then choose a Networking only cluster here. 6 00:00:25,630 --> 00:00:30,110 Click Next step, give this cluster a name of your choice, 7 00:00:30,110 --> 00:00:33,423 I'll name it goals-app but this name is totally up to you. 8 00:00:34,770 --> 00:00:38,950 Check Create VPC and keep all the defaults here, 9 00:00:38,950 --> 00:00:43,440 this ensures that AWS sets up private cloud 10 00:00:43,440 --> 00:00:46,600 for this cluster, for all of the containers in the cluster 11 00:00:47,460 --> 00:00:49,320 and click Create. 12 00:00:49,320 --> 00:00:51,330 Now this will create this cluster 13 00:00:51,330 --> 00:00:54,370 and the cluster is just the surrounding network 14 00:00:54,370 --> 00:00:56,690 for your containers thereafter. 15 00:00:56,690 --> 00:01:01,410 So let's wait for this cluster creation to finish, 16 00:01:01,410 --> 00:01:03,783 and this can take a couple of minutes. 17 00:01:05,960 --> 00:01:09,550 Once this finished click on View Cluster 18 00:01:09,550 --> 00:01:13,410 and now we need to add tasks and services here. 19 00:01:13,410 --> 00:01:15,480 Now services are based on tasks 20 00:01:15,480 --> 00:01:17,920 so we need to create a task first. 21 00:01:17,920 --> 00:01:20,740 And we do this under Task Definitions. 22 00:01:20,740 --> 00:01:24,110 Now I already got two task definitions here, 23 00:01:24,110 --> 00:01:26,580 we are going to create a new one though. 24 00:01:26,580 --> 00:01:30,370 So create a new task definition and choose Fargate here 25 00:01:30,370 --> 00:01:33,970 to have this serverless container execution environment 26 00:01:33,970 --> 00:01:37,020 where your containers can scale infinitely 27 00:01:37,020 --> 00:01:39,343 and you only pay for what you need. 28 00:01:40,240 --> 00:01:43,690 Click on Next step and give this task definition name 29 00:01:43,690 --> 00:01:45,833 and I'll name it goals. 30 00:01:48,810 --> 00:01:50,820 On task role you should have 31 00:01:50,820 --> 00:01:54,230 a ECS task execution role available, 32 00:01:54,230 --> 00:01:56,043 make sure you choose that one. 33 00:01:58,230 --> 00:02:01,930 If you don't have that, you should delete the entire cluster 34 00:02:01,930 --> 00:02:06,120 again and launch at first run wizard again and go through 35 00:02:06,120 --> 00:02:09,573 that instead because this will then create this role. 36 00:02:10,430 --> 00:02:14,280 Scroll down and choose the smallest sizes here 37 00:02:14,280 --> 00:02:18,440 for the memory and CPU to keep costs down. 38 00:02:18,440 --> 00:02:21,810 Of course for real apps you might wanna choose 39 00:02:21,810 --> 00:02:25,610 more powerful settings here to speed up your code execution, 40 00:02:25,610 --> 00:02:28,800 it depends on the app you're building and it will take time 41 00:02:28,800 --> 00:02:31,610 and experience to find out what you need here. 42 00:02:31,610 --> 00:02:35,360 And then on container definitions click on Add container. 43 00:02:35,360 --> 00:02:37,250 Here give your a container a name 44 00:02:37,250 --> 00:02:40,950 and I'll name it goals-node because that will be the backend 45 00:02:40,950 --> 00:02:44,670 or maybe therefore name it goals-backend container. 46 00:02:44,670 --> 00:02:49,590 And then use this repository here, 47 00:02:49,590 --> 00:02:53,123 this image name academind/goals-node in my case. 48 00:02:54,200 --> 00:02:58,470 On Port mappings make sure you map port 80 49 00:02:58,470 --> 00:03:02,660 because that is the port exposed by this application, 50 00:03:02,660 --> 00:03:07,660 we're listening on port 80 here in our node JS code. 51 00:03:11,410 --> 00:03:15,140 Ignore the health check but here under environment, 52 00:03:15,140 --> 00:03:17,263 I now do wanna change something. 53 00:03:18,130 --> 00:03:20,150 It is worth pointing out that 54 00:03:20,150 --> 00:03:22,900 in hour docker file for this backend image, 55 00:03:22,900 --> 00:03:27,900 we are actually executing npm start here as a final command. 56 00:03:28,160 --> 00:03:32,660 And that in turn will then utilize nodemon 57 00:03:33,530 --> 00:03:35,650 to start app js. 58 00:03:35,650 --> 00:03:40,650 Now I used nodemon locally to have this life reloading 59 00:03:40,890 --> 00:03:44,090 when we change something in our code. 60 00:03:44,090 --> 00:03:46,720 Now in production we won't have that, 61 00:03:46,720 --> 00:03:50,690 we also won't have to bind mounts to feed any live code 62 00:03:50,690 --> 00:03:51,970 into the running app, 63 00:03:51,970 --> 00:03:56,120 I talked about this early already, why we don't do that. 64 00:03:56,120 --> 00:03:59,400 So therefore we of course don't wanna use nodemon 65 00:03:59,400 --> 00:04:02,420 to execute our code here, we wanna use node. 66 00:04:02,420 --> 00:04:07,240 So therefore here on command I'll enter node app js. 67 00:04:07,240 --> 00:04:10,260 There should be a comma separated list here, 68 00:04:10,260 --> 00:04:13,200 cd hint here if nothing was entered 69 00:04:13,200 --> 00:04:17,930 and they offer node, app js will execute the app js file 70 00:04:17,930 --> 00:04:21,820 inside of the container with the node runtime. 71 00:04:21,820 --> 00:04:23,790 And that is a change we should make here 72 00:04:23,790 --> 00:04:28,790 and that also shows us how we again can leverage the exact 73 00:04:28,870 --> 00:04:32,640 same image and container during development and production 74 00:04:32,640 --> 00:04:36,350 and still find tune individual aspects 75 00:04:36,350 --> 00:04:39,810 to the different environments where we run the container, 76 00:04:39,810 --> 00:04:43,860 which is not a problem but actually a feature of docker. 77 00:04:43,860 --> 00:04:48,860 We still have the locked in code and delocked in runtime. 78 00:04:49,980 --> 00:04:53,120 We use node here, that's locked into the image. 79 00:04:53,120 --> 00:04:57,310 So we have that guarantee of having the same application 80 00:04:57,310 --> 00:05:01,100 and application environment in development and production, 81 00:05:01,100 --> 00:05:04,390 that main asset docker is all about. 82 00:05:04,390 --> 00:05:08,060 And still we can find tune individual aspects 83 00:05:08,060 --> 00:05:10,900 that need fine tuning like here 84 00:05:10,900 --> 00:05:13,380 where we use npm start locally 85 00:05:13,380 --> 00:05:16,663 and use node app js in production. 86 00:05:18,250 --> 00:05:21,160 Now we should specify the environment variables. 87 00:05:21,160 --> 00:05:23,250 And here I just noticed something, 88 00:05:23,250 --> 00:05:25,370 in the docker file we should of course 89 00:05:25,370 --> 00:05:28,380 also add this new environment variable, 90 00:05:28,380 --> 00:05:32,450 MONGODB URL environment variable which we're using 91 00:05:32,450 --> 00:05:34,770 in the MONGODB connection string now. 92 00:05:34,770 --> 00:05:38,330 So make sure you add this ENV definition 93 00:05:38,330 --> 00:05:42,760 in your docker file and then actually rebuild 94 00:05:42,760 --> 00:05:46,283 that docker file by running the docker build command again. 95 00:05:47,120 --> 00:05:49,480 And then of course also tag it again 96 00:05:49,480 --> 00:05:53,330 to the academind/goals-node image name in my case. 97 00:05:53,330 --> 00:05:56,100 So to that docker hub repository image, 98 00:05:56,100 --> 00:05:59,720 and then push updated image to ensure 99 00:05:59,720 --> 00:06:03,110 that your docker file lists all the environment variables 100 00:06:03,110 --> 00:06:06,083 which are getting used inside of the container of code. 101 00:06:07,070 --> 00:06:11,510 Now with that, back on AMS ECS we now can specify 102 00:06:11,510 --> 00:06:13,110 the environment variables here. 103 00:06:13,110 --> 00:06:16,730 And that's another important thing, locally 104 00:06:16,730 --> 00:06:20,590 when we run our container locally we use docker compose 105 00:06:20,590 --> 00:06:23,210 to run these multiple containers, 106 00:06:23,210 --> 00:06:28,110 and there we then of course used the env file configuration 107 00:06:28,110 --> 00:06:31,010 to point at an env file. 108 00:06:31,010 --> 00:06:34,420 Alternatively if we use to docker run command, 109 00:06:34,420 --> 00:06:39,280 there we also add an option for adding environment variables 110 00:06:39,280 --> 00:06:42,180 or pointing at an environment variable file. 111 00:06:42,180 --> 00:06:46,493 Here it was the env, the --env file option. 112 00:06:47,450 --> 00:06:49,720 Now we don't have an environment file here 113 00:06:49,720 --> 00:06:53,860 when we use AWS ECS but we can add our environment variables 114 00:06:53,860 --> 00:06:56,280 as key value pairs here. 115 00:06:56,280 --> 00:07:00,630 So therefore I now will go into the backend env file 116 00:07:00,630 --> 00:07:03,590 and grab the environment variables here 117 00:07:03,590 --> 00:07:05,940 and add them as keys. 118 00:07:05,940 --> 00:07:10,940 So MONGODB username, MONGODB password and MONGODB URL 119 00:07:13,180 --> 00:07:16,090 and then add values here. 120 00:07:16,090 --> 00:07:19,730 And I will use the same values as I did during development 121 00:07:19,730 --> 00:07:22,780 though in reality you might wanna consider choosing 122 00:07:22,780 --> 00:07:25,400 more complex usernames and passwords 123 00:07:25,400 --> 00:07:27,053 but I will go with these ones. 124 00:07:28,750 --> 00:07:31,367 But I will change my URL, 125 00:07:32,341 --> 00:07:37,300 I used MONGODB here in development because there thanks 126 00:07:37,300 --> 00:07:41,160 to docker compose and this network being created 127 00:07:41,160 --> 00:07:45,020 by docker which was a very useful feature here locally. 128 00:07:45,020 --> 00:07:48,770 We could use the container name as an address 129 00:07:48,770 --> 00:07:51,370 for requests being made from one container 130 00:07:51,370 --> 00:07:52,653 to another container. 131 00:07:53,879 --> 00:07:57,950 AWS ECS doesn't support this feature like this, 132 00:07:57,950 --> 00:08:00,570 but it does have another neat feature 133 00:08:00,570 --> 00:08:02,623 which is easy to set up and use. 134 00:08:03,530 --> 00:08:07,180 When we add multiple containers to the same task 135 00:08:07,180 --> 00:08:08,720 which we are going to do, 136 00:08:08,720 --> 00:08:11,250 we are going to add the MONGODB container 137 00:08:11,250 --> 00:08:15,110 to this same task as this container is part of, 138 00:08:15,110 --> 00:08:18,210 then these containers can communicate 139 00:08:18,210 --> 00:08:21,870 with each other under the local host key. 140 00:08:21,870 --> 00:08:25,250 So it basically emulates a local system 141 00:08:25,250 --> 00:08:28,973 where all containers are on the same host you could say. 142 00:08:29,870 --> 00:08:34,870 So here in production, in AWS ECS we can use local host 143 00:08:35,419 --> 00:08:38,320 to send the request to some other container 144 00:08:38,320 --> 00:08:41,250 which will be part of the same task. 145 00:08:41,250 --> 00:08:46,000 And therefore I use local host as a value for MONGODB URL 146 00:08:46,000 --> 00:08:51,000 in production, we use MONGODB as a value in development 147 00:08:52,200 --> 00:08:55,520 because in development local hosts wouldn't work. 148 00:08:55,520 --> 00:08:57,210 So that's again another difference 149 00:08:57,210 --> 00:09:00,710 but again we are implementing this such 150 00:09:00,710 --> 00:09:04,210 that our application code is not effected. 151 00:09:04,210 --> 00:09:06,850 There we just use an environment variable 152 00:09:06,850 --> 00:09:10,750 for getting the real address and then the value 153 00:09:10,750 --> 00:09:12,970 for that address depends simply 154 00:09:12,970 --> 00:09:17,150 on where we run this container, and that again ensures 155 00:09:17,150 --> 00:09:19,630 that what's in the container is locked in 156 00:09:19,630 --> 00:09:22,960 and works equally in all environments. 157 00:09:22,960 --> 00:09:26,980 And then we may just choose to run things differently 158 00:09:26,980 --> 00:09:28,823 in different environments. 159 00:09:30,170 --> 00:09:34,300 So with that, I added these free environment variables. 160 00:09:34,300 --> 00:09:38,090 We don't need to set up anything else here 161 00:09:38,090 --> 00:09:40,373 in the other parts. 162 00:09:42,150 --> 00:09:45,120 We also don't need to set up anything for storage 163 00:09:45,120 --> 00:09:48,800 and log in because the two volumes I did set up 164 00:09:48,800 --> 00:09:52,090 for my backend here, I removed the logs which we used 165 00:09:52,090 --> 00:09:55,070 in docker compose and therefore the two remaining volumes 166 00:09:55,070 --> 00:09:57,860 we have here are really just a bind mount 167 00:09:57,860 --> 00:10:00,350 and then an anonymous volume to ensure 168 00:10:00,350 --> 00:10:03,860 that the bind mount doesn't overwrite node modules. 169 00:10:03,860 --> 00:10:06,230 Since we're not using bind mounts in production, 170 00:10:06,230 --> 00:10:09,143 we therefore don't need to add any volume at all here. 171 00:10:10,220 --> 00:10:12,860 So therefore we can now click Add 172 00:10:12,860 --> 00:10:16,053 and with that we added our first container here. 173 00:10:16,960 --> 00:10:19,840 But of course that's not the only container we want here, 174 00:10:19,840 --> 00:10:22,270 after all we have an application 175 00:10:22,270 --> 00:10:27,120 that uses two containers here, the backend and MONGODB. 176 00:10:27,120 --> 00:10:29,813 So let's now add MONGODB as well.