checklist

[quarz12]

Service

• give better error messages if request fails due to missing profile, not "failed to load"
• roadmap
• improve embedding search?
• test coverage (esp. genai, client?)
• seed db in deployment
• remove sourcemaps
• azure logs
• k8 dashboards

Team and Process Compliance

• Team has 3 registered members with GitHub username, TUMonline login, and matriculation number documented.
• Each member has a primary subsystem (client/server/GenAI) and contributes to cross-subsystem integration.
• All work is done through feature branches and pull requests.
• No normal direct commits to main.
• Peer review and approval happen before each merge.
• Contribution traceability is visible via commits, PR authorship, reviews, and infra work.
• Planning/questions/feedback are tracked in official Artemis channels.
  System Architecture and Stack Compliance
• Mono-repo contains client, server, GenAI, deployment, CI/CD, and docs.
• Client is implemented with an allowed framework and communicates via REST.
• Server is Spring Boot and contains at least 3 distinct microservices.
• Database is persistent and schema is documented.
• GenAI service is a separate Python microservice with defined interface.
• GenAI provides a real user-facing capability (not only technical presence).
• OpenAPI/Swagger documentation exists and is exposed.

Containerization and Local Reproducibility

• Each component has its own Dockerfile (client, server services, GenAI, DB).
• Full system runs end-to-end via compose setup.
• Local startup is possible in 3 commands or fewer.
• Setup instructions are reproducible for a new user without hidden manual steps.

Kubernetes and Deployment Compliance

• Deployment manifests/charts exist for Kubernetes.
• CD deploys automatically to Kubernetes on merge to main.
• Deployment works on course infra (Rancher) and Azure.
• Environment-specific config is externalized (env vars, secrets, config files).
• No hardcoded credentials/tokens/environment constants in source code.
• A stable deployed URL is available for tutor interaction.

CI/CD Compliance

• GitHub Actions pipeline runs on every PR.
• CI builds all relevant services.
• CI runs automated tests.
• CI includes linting/static analysis where appropriate.
• CI fails reliably on broken code/tests.

Observability Compliance

• Prometheus is integrated for metrics collection.
• Metrics include at least request count, latency, and error rate.
• Grafana dashboards visualize meaningful system behavior.
• Grafana dashboards are exported as .json files for submission.
• At least one meaningful alert rule is configured and documented.

Testing Compliance

• Unit tests cover critical server logic.
• Unit tests cover relevant GenAI logic.
• Client tests cover key user workflows/interactions.
• All tests run automatically in CI.

Engineering Artifacts and Documentation Compliance

• High-level architecture description is available and matches implementation.
• Mandatory diagrams are present: Subsystem Decomposition, Use Case, Analysis Object Model.
• API documentation and Swagger/OpenAPI references are in README/docs.
• README documents setup, architecture, API docs, CI/CD, monitoring, and responsibilities.
• Monitoring config, dashboards, and alerts are versioned in the repo.
• Testing instructions are documented.

Presentation and Oral Exam Readiness

• Team can demonstrate working end-to-end system live.
• Team can explain architecture, pipeline, trade-offs, and operational aspects.
• Each member can clearly explain and defend their own subsystem/artifact.
• Evidence of monthly progress/checkpoint improvements is prepared.

Failure Guard (Must Be True)

• Contributions are transparently documented (GitHub + Artemis evidence).
• Every member can explain their own subsystem during presentation/exam.
• End-to-end system works in a real demo.