Application Service
Last updated
Was this helpful?
Last updated
Was this helpful?
EC2 instances run in ECS Cluster
Allow to manage the underlying infrastructure
Make API calls to ECS service
Send log to cloudwatch logs
Based on the task definition (the image that want to run)
Increase numbers of tasks for scaling
Supports auto-scaling, but it requires manual configuration. You need to define and manage scaling policies based on metrics like CPU utilization or request count. ECS gives you more control over the scaling behavior and allows you to customize it according to your application's needs.
Both are used docker container as the smallest unit, task (ECS) & pod (EKS)
Both are container orchestration service
Highly available and scalable. Also supports both manual configuration and automated scaling without running out of control
EKS is a managed K8s service you can run on any infrastructure — cloud or on-premises, increasing portability between vendors and lowering vendor-lock in
Applying namespace , ingress, service , etc is needed for EKS, but ECS has simple API eliminates complexity while using Route 53 , Elastic Load Balancer , IAM and Cloud Watch for logging
Fully managed service that make it easy to deploy
Suitable for Container-based application
Automatically scaling, highly available, load balacing are included
Its nature is actually event listener, listen for SQS message, HTTP request, ...
Paid per request and compute time
Free tiers of 1000000 requests and 40000 GBs
Memory Allocation: 128MB - 10GB
Maximum execution time: 900s
Max. env variable: 4KB
Concurrent execution: 1000
Compressed deployment size: 50MB
Uncompressed deployment size: 250MB
Max. 10 messages from queue can handled for each invocation
RDS PostgresDB and Aurora MYSQL can send event notification about DB instance itself (created, stopped, ...) to trigger
Suitable for event-driven application, such as report function
Similar with firebase, supabase
providing whole set of function as amplify SDK , but also including CI/CD
Amplify can be used in frontend side to conduct CRUD, authentication logic
Suitable for fast "full stack" development
Used to change the request and response
On CloudFront Function, It can create token for authentication and authorization
There are 2 endpoint types
Edge-Optimized (default): for global client, request is routed through cloud front edge location, the gateway lives in 1 region
Regional: For regional client
Private: Need to access from VPC
As an inter-communication channel between different services within the application.
Can be triggered by action of aws service to react to trigger the logic in the destination
Can be triggered by schedule job
Introduction
Provide users' identity for accessing application or aws resources
Integrated with API Gateway or load balancer for authentication to application
To grant the temporary access for aws resources
The IAM policy applied to access is defined