AWS App Runner

Overview

AWS App Runner is a fully managed service that makes it easy to deploy web applications and APIs at scale. You need no infrastructure knowledge.

1. Start with your source code or a Docker container image.
2. Configure settings: vCPUs, memory, auto scaling, health checks.
3. App Runner builds and deploys the web app or API.
4. Access it right away through a URL.

Features

• Automatic scaling, highly available, load balancing, encryption
• Access to your VPC: connect to databases, caches and message queue services

Hands-On Walkthrough

The instructor notes this is not in the free tier: you pay for vCPU and per GB.

1. Create an App Runner service. Source: Container registry (ECR for private images, ECR Public for public images) or source code (e.g. from GitHub).
2. Use a public image: from the ECR Public Gallery copy the httpd image address and paste it.
3. Deployment setting: manual or automatic (automatic available with a source code repo or ECR). Manual here.
4. Service name DemoHTTP; 1 vCPU and 2 GB (the lowest available). Optionally environment variables; port 80 (the image's port).
5. Auto scaling default: 1 to 25 instances with 100 concurrent requests per instance; a custom configuration can be added.
6. Health checks left as is. Security: optional instance role. Networking: public access (or custom VPC). Optional tracing with AWS X-Ray. Defaults, next.
7. Review and Create and deploy. Behind the scenes App Runner deploys an auto scaling group, scaling triggers, target containers, a load balancer and a domain (the speaker says "most likely" for some of these).
8. One interface gives logs, activity, metrics, observability, configuration and custom domain connection.
9. The default domain shows "It works", the message for a correctly deployed HTTPD. The event log shows deployment and health checks. Many concurrent requests would scale it automatically.
10. Delete the service by typing delete.

Questions & Answers