Blue/green deployments?
Elastic Beanstalk?
## Cloud Practitioner:
### Scalability: Use the resources you need being capable of increasing or
decommission them depending on its demand. AWS service that autoscales: Amazon EC2
Auto Scaling.
It automatically autoscales your instances. It can scale in various
approaches:
*Dynamic: Responds to actual demand.
*Predictive: Responds to predicted demand.
*Forms of scaling:
*Scale up (expansión vertical): Adding more power to the machines. I.e.
adding more RAM, more CPU cores...
*Scale out (expansión horizontal):
Amazon EC2 Autoscaling Group: You can define the minimum instances for your
group (the minimum instances your application needs), the desired amount of
instances (the number of instances that will work in your app), and the maximum
instances your app can use depending on the demand. If you do not select a desired
amount, the working amount will be the minimum number of instances needed.
The service that manages the scaling of instances is Amazon Elastic Load
Balancing.
### Types of Architecture:
*Tightly Coupled Architecture: The components that interact with each other
are dependent and must be present and functioning at the same time. If a single
component fails, it causes issues for other components. It's characteristic of
monolitic applications.
*Loosely Coupled Architecture: The components can function independently. If
a single component fails, it won't cause cascade failures. Is the architecture of
AWS. I.e. One app sends messages to a buffer/queue that stores them until the
destination app is up. This service is called Amazon Simple Queue Service (SQS).
It's characteristic of microservices.
### Serverless Computing: You don't need to manage the servers that your
application needs. Also the scalation is automatic. E.g.: Amazon Lambda. You are
charged for the time your code is running.
AWS Fargate: Serverless compute engine for containers.
### AWS Infraestructure: AWS has Availability Zones (single data center or a group
of them) that group within a Region. AZs are separated physically from themselves.
One instance runs on one AZ, that's why you should have at least two instances in
two different AZs (so if one crashes, your app will still be available).
The data is available for clients between regions with Amazon CloudFront.
Between regions are "AWS Edge Locations" which CloudFront uses to send data with
low latency.
### Dictionary
* VPC (Virtual Private Cloud):
## PREGUNTAS EX.
Obtain public IP address of a service:
From the instance metadata service associated with the EC2 instance.
