Per-package isolated Postgres schema with migrations. Local benchmark: 12.87s sequential → 4.51s parallel = 2.85x. ADR-0025 status to Implemented. CI uses BDD_SCHEMA_ISOLATION=true.
Co-authored-by: Gabriel Radureau <arcodange@gmail.com>
Co-committed-by: Gabriel Radureau <arcodange@gmail.com>
Extends cache service to /api/v1/greet (per-name 60s) and adds POST /api/admin/cache/flush. ~95% Mistral autonomous via ICM workspace, trainer finalized commit/PR (test scaffold did not compile).
Co-authored-by: Gabriel Radureau <arcodange@gmail.com>
Co-committed-by: Gabriel Radureau <arcodange@gmail.com>
Phase 1 part 2 of ADR-0022 (companion to PR #22 rate-limit). In-memory cache service via go-cache, used by /api/version (60s TTL).
6/6 unit tests pass. ~95% Mistral autonomous via ICM workspace, cost €2.50 stages 01-02 (50% reduction vs T5 thanks to pre-extracted snippets in shared/).
Co-authored-by: Gabriel Radureau <arcodange@gmail.com>
Co-committed-by: Gabriel Radureau <arcodange@gmail.com>
Adds Kubernetes-style /api/healthz endpoint with status/version/uptime_seconds/timestamp.
Non-breaking — /api/health preserved. Includes unit test (passes locally) and BDD scenario (validated by CI).
Généré ~95% en autonomie par Mistral Vibe via workspace ICM ~/Work/Vibe/workspaces/healthz-feature/.
Co-authored-by: Gabriel Radureau <arcodange@gmail.com>
Co-committed-by: Gabriel Radureau <arcodange@gmail.com>
## Summary
Closes#15
When `logging.json: true` (or `DLC_LOGGING_JSON=true`), the logger was unconditionally initialised to console/text format at the top of `LoadConfig()`, so early log lines — most visibly **"Config file loaded"** — were always written as human-readable text regardless of configuration.
## Root cause
Classic chicken-and-egg: the format flag lives inside the config that is being loaded. The format-switch block only ran *after* `v.Unmarshal()`, too late for the config-file log.
## Changes
### `pkg/config/config.go`
- Add `peekJSONLogging()`: resolves the JSON flag **before** any log is emitted by (1) checking `DLC_LOGGING_JSON` directly via `os.Getenv`, then (2) doing a minimal throwaway Viper pre-read of the config file for the `logging.json` key. This mirrors Viper's own priority order without parsing the full config twice.
- Apply the resolved format immediately and emit **"Logging configured"** as the very first log line.
- Remove the now-redundant format-switch block that ran after `Unmarshal()`.
### `scripts/start-server.sh`, `test-graceful-shutdown.sh`, `test-opentelemetry.sh`
- Replace hardcoded `PROJECT_DIR` path with a dynamic `SCRIPTS_DIR=$(dirname $(realpath ${BASH_SOURCE[0]}))` derivation so scripts work from any worktree or clone location.
## Test plan
- [x] `go test ./pkg/...` — all pass
- [x] `scripts/test-graceful-shutdown.sh` — all JSON valid, all startup logs present
- [x] Manual smoke test: first line is `{"level":"info",...,"message":"Logging configured"}`, every line is valid JSON
Reviewed-on: #16
Co-authored-by: Gabriel Radureau <arcodange@gmail.com>
Co-committed-by: Gabriel Radureau <arcodange@gmail.com>
- Add swag fmt to git pre-commit hook and CI/CD pipeline
- Create comprehensive CONTRIBUTING.md guide with AI section
- Update ADR-0013 with swag fmt documentation
- Fix swagger generation to include all endpoints
- Improve local testing scripts and workflows
- Update Dockerfile for better swagger handling
- Fix CI/CD workflow file references
- Add Version field to telemetry.Setup struct
- Include service.version attribute in OpenTelemetry resource
- Pass version.Short() from server to telemetry initialization
- Follows OpenTelemetry semantic conventions for better observability
- Enables version tracking in production traces
Generated by Mistral Vibe.
Co-Authored-By: Mistral Vibe <vibe@mistral.ai>
- Add ADR-0012 documenting the decision to format only staged Go files
- Update ADR README.md with new entry
- Document rationale, alternatives, and verification results
- Include future considerations for monitoring and CI/CD integration
Generated by Mistral Vibe.
Co-Authored-By: Mistral Vibe <vibe@mistral.ai>
- Added /api/v2/greet POST endpoint with JSON request/response
- Implemented ServiceV2 with Hello my friend <name>! greeting format
- Added api.v2_enabled feature flag (default: false)
- Extended BDD tests to cover v2 scenarios
- Maintained full backward compatibility with v1 API
- Added DLC_API_V2_ENABLED environment variable support
- Created ADR 0010-api-v2-feature-flag.md
- Updated configuration system to support API versioning
The time.Sleep(100ms) after OpenTelemetry shutdown was removed as it's
not necessary for proper log writing. Zerolog automatically flushes logs
on program exit, and the remaining 1-second sleep for readiness
propagation is sufficient.
Tested that all shutdown logs are properly written without the sleep.
- Simplify cmd/server/main.go from 158 to 57 lines (64% reduction)
- Move HTTP server creation, graceful shutdown, and context management to pkg/server
- Add Run() method to encapsulate server lifecycle management
- Maintain all existing functionality and OpenTelemetry integration
- Improve separation of concerns and code organization
This refactoring makes cmd/server/main.go a thin entrypoint while moving all server implementation details to the pkg/server package, following Go best practices for package organization.
Add server and greet binaries to .gitignore
Implement readiness endpoint (/api/ready) that returns:
- {"ready":true} (HTTP 200) during normal operation
- {"ready":false} (HTTP 503) during graceful shutdown
Key changes:
- Added readiness context to control readiness state
- Modified server.NewServer() to accept readiness context
- Implemented handleReadiness() with context-aware logic
- Updated cmd/server/main.go to manage readiness state
- Readiness set to false when shutdown signal received
- Updated test script to validate readiness behavior
- Added comprehensive documentation for readiness endpoint
This allows Kubernetes/service meshes to stop routing traffic
to the pod during graceful shutdown while allowing existing
requests to complete. Health endpoint continues to return
happy status during shutdown for proper orchestration.
- Added signal.NotifyContext for modern signal handling in cmd/server/main.go
- Implemented BaseContext for proper context propagation to HTTP handlers
- Added readiness drain delay before shutdown for graceful degradation
- Fixed PID detection in start-server.sh to target actual server process
- Added Logging.JSON configuration option with DLC_LOGGING_JSON environment variable
- Created comprehensive test script that validates entire server lifecycle
- Updated documentation with JSON logging configuration examples
- All shutdown logs now appear correctly in JSON format
- Server terminates gracefully on SIGTERM with proper log flushing
The graceful shutdown implementation follows VictoriaMetrics best practices:
1. Catches termination signals (SIGTERM, SIGINT)
2. Stops accepting new requests but allows ongoing requests to complete
3. Waits for active requests to finish within configured timeout
4. Releases resources and performs cleanup
5. Logs all shutdown steps for observability
Test script validates:
- Server startup and API functionality
- Graceful shutdown sequence
- JSON log format validation
- Complete log sequence verification
- Proper signal handling and context propagation
Generated by Mistral Vibe.
Co-Authored-By: Mistral Vibe <vibe@mistral.ai>
\n- Add scripts/start-server.sh for easy server management\n- Add DLC_NO_COLOR environment variable support for clean logs\n- Update server.go to handle color/no-color logging\n- Add server.log and server.pid to .gitignore\n- Update README.md and AGENTS.md with server control documentation\n- Test all server control functions (start, stop, restart, status, logs, test)
- Added context-aware server initialization in cmd/server/main.go
- Implemented graceful shutdown handling with SIGINT/SIGTERM signals
- Added 30-second shutdown timeout for active connections
- Updated Greet service to use context.Context as first parameter
- Enhanced Zerolog integration with Trace level logging
- Added context-aware logging in Greet function calls
- Fixed route structure to use /api/v1/greet/* prefix
- Updated all handlers and tests to use context
- Comprehensive AGENTS.md documentation with verified commands
- Added server context management architecture section
- Updated API endpoint documentation with working examples
Changes:
- cmd/server/main.go: Complete rewrite with context and graceful shutdown
- pkg/greet/greet.go: Added context parameter and trace logging
- pkg/greet/api_v1.go: Updated interface and handlers for context
- pkg/greet/greet_test.go: Updated tests to use context
- cmd/greet/main.go: Updated CLI to use context
- pkg/server/server.go: Trace level config and context logging
- AGENTS.md: Comprehensive documentation update
- go.mod/go.sum: Added Zerolog dependency
All tests passing, server working with graceful shutdown verified.
- Add pkg/greet/api_v1.go with ApiV1Greet interface and handler
- Implement Greeter interface for dependency injection
- Return JSON responses with proper Content-Type headers
- Move greet service instantiation to server package
- Separate v1 route registration into registerApiV1Routes function
- Health endpoint at /api level, greet endpoints at /api/v1/greet
- Simplify server entrypoint by removing external dependencies
