Infrastructure-as-code

clients

• to create an aws service, we must first create a client for each corresponding service

monitor ec2 instance

• to check if the script was installed correctly on created ec2 instance
• connect to the ec2 instance:
• click on the instance id in "EC2 Dashboard"
• click on "connect" button
• in tab "EC2 instance connect"
• connection type is "Connect using EC2 Instance Connect"
• click on "Connect"
• by default the connection is in user folder, you want to go to root folder
• enter command "cd .." and log content with command "ls" until you go to root folder and find all files
• when you find "app" folder, go inside using command "cd app"
• make sure all files like "main.py" and and environment "venv" are created
• in another tab, go to instance's details and copy value of "Public IPv4 address" and paste it in a browser tab to check if app is running

debug app error

• first deactivate the background run
• use command [ ps aux | grep "python3 main.py" ] to find the process running in background
• use command "kill process_id"
• manually run the app using "python3 main.py"
• use browser to ping API and check error logs in the connect terminal

aws listeners

• we attach listeners to the load balancer to tell how to route the traffic

target groups

• we define target group and put resource associated with them
• the listener can then forward traffic to that group

flask app

• for routing inside the application, we need to define as many routing as the number of cluster
• if 2 cluster then /cluster1 and /cluster2
• else the load balancer rules cannot be define automatically

delete resource order

• sometimes will have to wait a bit for other resources to detect that their user resources are deleted
• delete all running instances
• delete the load balancer
• delete all target groups (found when click on a random running instance, check left side bar and scroll down)
• delete the security group

MetricName

RequestCount: This metric represents the total number of requests handled by the load balancer. HealthyHostCount: This metric counts the number of healthy instances behind the load balancer. UnHealthyHostCount: This metric counts the number of unhealthy instances. Latency: Measures the time it takes to complete a request.

Dimensions

• specific on a metric
• The availability region
• The target group
• the resource

uncomment json print when needed

• to check how a filter works uncomment the print json to see the full object

graph

• if folder "graphs" is not created, created else code will crash