Chore: Enable Nullable, Analyzers, and Structured Observability

[webreidi]

Goal

Raise code quality bar and improve runtime diagnosability by enabling nullable reference types, adding analyzers, enforcing warnings-as-errors, and instrumenting structured logging + tracing for critical game flows.

Motivation

• Earlier detection of null bugs & API misuse
• Consistent style / code hygiene reduces PR review friction
• Observability (logs + traces) accelerates debugging of gameplay issues, latency, hot paths

Quality Gate Enhancements

1. Enable <Nullable>enable</Nullable> in every csproj
2. Add <TreatWarningsAsErrors>true</TreatWarningsAsErrors> (optionally configurable via Directory.Build.props)
3. Add Microsoft.CodeAnalysis.NetAnalyzers (latest stable) with rule level medium severity; include editorconfig baseline if noise high
4. Add StyleCop.Analyzers (optional) or rely on .editorconfig for minimal style safety
5. Configure CI (GitHub Actions future) to run dotnet build + filtered tests

Observability Additions

• Structured logging (Microsoft.Extensions.Logging) in GameService: events for GameStarted, GuessSubmitted, GameCompleted with contextual properties (GameId, AttemptIndex, WordLength, Result)
• Exception logging with reduced noise (log domain validation at Information or Debug, unexpected at Error)
• OpenTelemetry instrumentation (if Aspire not already auto-wiring):
  • Add ActivitySource Codele.Gameplay
  • Span per guess evaluation with attributes: codele.game.id, codele.guess.length, codele.guess.result (win/continue)
• Optional metrics (if instrumentation libs added later): counter codele.guesses.total, histogram codele.game.duration.seconds

Implementation Plan

1. Create/modify Directory.Build.props to centrally enable Nullable + WarningsAsErrors; reference analyzers package
2. Add .editorconfig with key rules (nullable, IDE, formatting)
3. Update each project csproj to remove redundant settings; add analyzers package if not global
4. Introduce Instrumentation static class exposing ActivitySource
5. Inject ActivitySource into services (or use static) & wrap guess evaluation in activity
6. Add structured logging statements (Ensure no PII / target word withheld until game completes; only log masked hash before completion)
7. Add unit test asserting logger called for GameCompleted (can use a fake logger) – lightweight
8. Update README with new quality gates + how to suppress intentionally (pragma + justification)

Acceptance Criteria

• Build fails on new nullable warnings introduced intentionally
• Activity exported (can be verified via debug logging or future OTLP exporter)
• Logging shows contextual properties per guess in local run
• README documents process

Out of Scope

• Full metrics pipeline export
• External log shipping config

Risk & Mitigation

• Analyzer noise -> Keep initial rules moderate; create baseline suppression file if needed

Definition of Done

Merged PR with green tests, updated docs, analyzers active (visible in IDE), and logged gameplay events.

Estimated Effort

Small-Medium (3-4h depending on analyzer tuning).

Depends on: Issue #8 (for domain service centralization makes instrumentation cleaner).