A personal project to build and operate a platform on Microsoft Azure: infrastructure as code, a Kubernetes cluster, delivery pipelines, monitoring, incident handling and cost control.

The goal is to practice the full SRE and DevOps loop on one system: build it, deploy it, measure it, break it and fix it.

The repository is at github.com/Franklindot04/production-platform-Azure. These pages are the files of its docs/ folder.

Design and plan

• Architecture: components, layers, environments, identity, cost and open questions.
• Plan: the plan for the remaining epics and the live sessions.
• Project history: how the repository was built and why its history is short.
• Saved plans: the saved Terraform plans of the dev roots.

Delivery and security

• Workflows: the GitHub Actions workflows.
• Security: roles and secret rules.

Reliability

• SLOs: the SLIs, the targets and the burn-rate alerts.
• Incident process: severity levels and what to do at each step.
• Runbooks: one procedure per known failure, with the status of each.
• Postmortems: the rules, the template and the written postmortems.

Live sessions

• Session 1: the plan.
• Session 1 results
• Session 2: the plan.
• Session 2 results

Changes

• Changelog: what is on main.
• Release notes: no release has been published.

Decision records

• ADR 0001: Template
• ADR 0002: Record decisions as ADRs
• ADR 0003: Choose Azure region and resource naming convention
• ADR 0004: Use Azure Storage as the Terraform state backend
• ADR 0005: Choose the AKS networking model
• ADR 0006: Choose the GitOps tool
• ADR 0007: Authenticate GitHub Actions to Azure with OIDC
• ADR 0008: Choose the metrics and dashboard stack
• ADR 0009: Set cost guardrails and teardown policy
• ADR 0010: Separate environments by resource group
• ADR 0011: Design the session network
• ADR 0012: Plan the first AKS cluster
• ADR 0013: Use GitHub Container Registry for images
• ADR 0014: Write the sample service in Go
• ADR 0015: Use Azure network policy
• ADR 0016: Choose the ingress
• ADR 0017: Bootstrap Argo CD from the session script
• ADR 0018: Update the image tag with a pull request
• ADR 0019: Choose the trace backend and Grafana hosting
• ADR 0020: Choose policy and scanning tools
• ADR 0021: Sign in to the cluster with Entra ID
• ADR 0022: Define the persistent layer and the pull request plan identity
• ADR 0023: Serve metrics on the admin port
• ADR 0024: Enable managed Prometheus on the cluster
• ADR 0025: Send traces with the Azure Monitor exporter and a connection string
• ADR 0026: Trace the sample service with the OpenTelemetry SDK
• ADR 0027: Keep dashboard queries as files under monitoring/
• ADR 0028: Set SLOs for the sample service and alert on burn rate
• ADR 0029: Give the sample service access to Key Vault with workload identity
• ADR 0030: Set the incident process and write postmortems