DevOps On AWS and AI - Real Projects - All In One

Install docker on a jupyter server using a script to automate the setup, then verify the installation by checking the docker version and granting docker user permissions.

Learn to understand and construct dockerfiles for Python, Node.js, and Java apps by stepwise copying dependencies, installing packages, running unit tests, and configuring multi-stage builds and ports.

Learn how to use docker compose to build images, configure services like mongodb and mysql, set up volumes, networks, and environments, and run step-by-step deployments.

Set up ECR to store a versioned image on AWS, configure permissions, build and test the image, and verify login success.

Create a GKE cluster and a VPC with Terraform, then verify the deployment. Check the VPC dashboard, subnets, root tables, internet gateway, cluster resources, IAM, and routes like DksMathRoute.

Connect to an EKS cluster by running setup commands and use kubectl to inspect nodes, pods, deployments, services, ingress, and other resources in the cluster.

Set up MongoDB Atlas free tier, create a database and collection, configure a database user with a username and password, and save the connection string.

Connect to MongoDB Cloud, manage sequences, and insert documents with 5 and 2 stars; configure IP allowlists to run the app.

Push a CI/CD workflow to the project and run it with GitHub Actions, push images to ECR, and deploy to EKS to verify the running app.

Create a gateway and http route, then access the app through the load balancer domain using the address from kubectl get gateway, and verify status across versions 1 to 3.

Set up a service and its route for a single version, then create three preview versions of the service to observe the results.

Configure an http route to serve any reviews version and switch between version 1, 2, and 3. Observe logs to verify the active version and ensure correct routing.

This lecture demonstrates horizontal autoscaling and monitoring, simulating traffic to trigger replica scaling based on CTQ target utilization, with a 60-second stabilization window.

Build and edit Grafana dashboards to monitor real-time cpu usage and application performance, using simulated traffic to visualize bot cpu metrics across configurable time windows.

Set up a Grafana to Slack alert workflow by creating a cdqbot app with an incoming webhook, linking it to the cdqmonitoring channel, and testing notifications.

Set up a domain namespace in Cloudflare, connect your domain, and enable SSL, choosing the Korean namespace plan to manage the domain. Expect time for certification, from minutes to hours.

Install cert-manager to issue and manage SSL certificates, securing your domain, data, and traffic, then enable guest-wide APIs to route traffic and advance to the next step.

Scale the cluster to three nodes by applying Terraform to add compute and pods, using a horizontal autoscaler for testing across platforms on Nextstack.