docs: update README for Phase 3 infrastructure completion

Update README to reflect completed Phase 3 infrastructure layer:
- Documented ModbusRelayController, MockRelayController, SqliteRelayLabelRepository, and HealthMonitor implementations
- Added testing coverage details (20+ tests across infrastructure components)
- Updated architecture diagrams and project structure
- Changed task reference to tasks.org format
- Updated dependency list with production infrastructure dependencies

Ref: Phase 3 tasks in specs/001-modbus-relay-control/tasks.org

README.md

@@ -62,33 +62,59 @@ STA will provide a modern web interface for controlling Modbus-compatible relay
 - RelayController and RelayLabelRepository trait definitions
 - Complete separation from infrastructure concerns (hexagonal architecture)
 
-### Planned - Phases 3-8
+### Phase 3 Complete - Infrastructure Layer
-- 📋 Modbus TCP client with tokio-modbus (Phase 3)
+- ✅ T028-T029: MockRelayController tests and implementation
-- 📋 Mock controller for testing (Phase 3)
+- ✅ T030: RelayController trait with async methods (read_state, write_state, read_all, write_all)
-- 📋 Health monitoring service (Phase 3)
+- ✅ T031: ControllerError enum (ConnectionError, Timeout, ModbusException, InvalidRelayId)
+- ✅ T032: MockRelayController comprehensive tests (6 tests)
+- ✅ T025a-f: ModbusRelayController implementation (decomposed):
+  - Connection setup with tokio-modbus
+  - Timeout-wrapped read_coils and write_single_coil helpers
+  - RelayController trait implementation
+- ✅ T034: Integration test with real hardware (uses #[ignore] attribute)
+- ✅ T035-T036: RelayLabelRepository trait and SQLite implementation
+- ✅ T037-T038: MockRelayLabelRepository for testing
+- ✅ T039-T040: HealthMonitor service with state tracking
+
+#### Key Infrastructure Features Implemented
+- **ModbusRelayController**: Thread-safe Modbus TCP client with timeout handling
+  - Uses `Arc<Mutex<Context>>` for concurrent access
+  - Native Modbus TCP protocol (MBAP header, no CRC16)
+  - Configurable timeout with `tokio::time::timeout`
+- **MockRelayController**: In-memory testing without hardware
+  - Uses `Arc<Mutex<HashMap<RelayId, RelayState>>>` for state
+  - Optional timeout simulation for error handling tests
+- **SqliteRelayLabelRepository**: Compile-time verified SQL queries
+  - Automatic migrations via SQLx
+  - In-memory mode for testing
+- **HealthMonitor**: State machine for health tracking
+  - Healthy -> Degraded -> Unhealthy transitions
+  - Recovery on successful operations
+
+### Planned - Phases 4-8
 - 📋 US1: Monitor & toggle relay states - MVP (Phase 4)
 - 📋 US2: Bulk relay controls (Phase 5)
 - 📋 US3: Health status display (Phase 6)
 - 📋 US4: Relay labeling (Phase 7)
 - 📋 Production deployment (Phase 8)
 
-See [tasks.md](specs/001-modbus-relay-control/tasks.md) for detailed implementation roadmap (102 tasks across 9 phases).
+See [tasks.org](specs/001-modbus-relay-control/tasks.org) for detailed implementation roadmap.
 
 ## Architecture
 
 **Current:**
-- **Backend**: Rust 2024 with Poem web framework
+- **Backend**: Rust 2024 with Poem web framework (hexagonal architecture)
 - **Configuration**: YAML-based with environment variable overrides
 - **API**: RESTful HTTP with OpenAPI documentation
 - **CORS**: Production-ready configurable middleware with security validation
-- **Middleware Chain**: Rate Limiting → CORS → Data injection
+- **Middleware Chain**: Rate Limiting -> CORS -> Data injection
+- **Modbus Integration**: tokio-modbus for Modbus TCP communication
+- **Persistence**: SQLite for relay labels with compile-time SQL verification
 
 **Planned:**
-- **Modbus Integration**: tokio-modbus for Modbus TCP communication
 - **Frontend**: Vue 3 with TypeScript
 - **Deployment**: Backend on Raspberry Pi, frontend on Cloudflare Pages
 - **Access**: Traefik reverse proxy with Authelia authentication
-- **Persistence**: SQLite for relay labels and configuration
 
 ## Quick Start
 
@@ -205,48 +231,65 @@ sta/                          # Repository root
 │   │   ├── lib.rs           - Library entry point
 │   │   ├── main.rs          - Binary entry point
 │   │   ├── startup.rs       - Application builder and server config
-│   │   ├── settings/        - Configuration module
-│   │   │   ├── mod.rs       - Settings aggregation
-│   │   │   └── cors.rs      - CORS configuration (NEW in Phase 0.5)
 │   │   ├── telemetry.rs     - Logging and tracing setup
-│   │   ├── domain/          - Business logic (NEW in Phase 2)
+│   │   │
-│   │   │   ├── relay/       - Relay domain types, entity, and traits
+│   │   ├── domain/          - Business logic layer (Phase 2)
+│   │   │   ├── relay/       - Relay domain aggregate
 │   │   │   │   ├── types/   - RelayId, RelayState, RelayLabel newtypes
-│   │   │   │   ├── entity.rs    - Relay aggregate
+│   │   │   │   ├── entity.rs    - Relay aggregate with state control
-│   │   │   │   ├── controller.rs - RelayController trait
+│   │   │   │   ├── controller.rs - RelayController trait & ControllerError
-│   │   │   │   └── repository.rs - RelayLabelRepository trait
+│   │   │   │   └── repository/   - RelayLabelRepository trait
 │   │   │   ├── modbus.rs    - ModbusAddress type with conversion
 │   │   │   └── health.rs    - HealthStatus state machine
-│   │   ├── application/     - Use cases (planned Phase 3-4)
+│   │   │
+│   │   ├── application/     - Use cases and orchestration (Phase 3)
+│   │   │   └── health/      - Health monitoring service
+│   │   │       └── health_monitor.rs - HealthMonitor with state tracking
+│   │   │
 │   │   ├── infrastructure/  - External integrations (Phase 3)
-│   │   │   └── persistence/ - SQLite repository implementation
+│   │   │   ├── modbus/      - Modbus TCP communication
+│   │   │   │   ├── client.rs       - ModbusRelayController (real hardware)
+│   │   │   │   ├── client_test.rs  - Hardware integration tests
+│   │   │   │   └── mock_controller.rs - MockRelayController for testing
+│   │   │   └── persistence/ - Database layer
+│   │   │       ├── entities/       - Database record types
+│   │   │       ├── sqlite_repository.rs - SqliteRelayLabelRepository
+│   │   │       └── label_repository.rs  - MockRelayLabelRepository
+│   │   │
 │   │   ├── presentation/    - API layer (planned Phase 4)
+│   │   ├── settings/        - Configuration module
+│   │   │   ├── mod.rs       - Settings aggregation
+│   │   │   └── cors.rs      - CORS configuration
 │   │   ├── route/           - HTTP endpoint handlers
 │   │   │   ├── health.rs    - Health check endpoints
 │   │   │   └── meta.rs      - Application metadata
 │   │   └── middleware/      - Custom middleware
 │   │       └── rate_limit.rs
+│   │
 │   ├── settings/            - YAML configuration files
 │   │   ├── base.yaml        - Base configuration
-│   │   ├── development.yaml - Development overrides (NEW in Phase 0.5)
+│   │   ├── development.yaml - Development overrides
-│   │   └── production.yaml  - Production overrides (NEW in Phase 0.5)
+│   │   └── production.yaml  - Production overrides
 │   └── tests/               - Integration tests
-│       └── cors_test.rs     - CORS integration tests (NEW in Phase 0.5)
+│       └── cors_test.rs     - CORS integration tests
+│
+├── migrations/              - SQLx database migrations
 ├── src/                     # Frontend source (Vue/TypeScript)
 │   └── api/                 - Type-safe API client
 ├── docs/                    # Project documentation
 │   ├── cors-configuration.md    - CORS setup guide
-│   ├── domain-layer.md          - Domain layer architecture (NEW in Phase 2)
+│   ├── domain-layer.md          - Domain layer architecture
 │   └── Modbus_POE_ETH_Relay.md  - Hardware documentation
 ├── specs/                   # Feature specifications
 │   ├── constitution.md      - Architectural principles
 │   └── 001-modbus-relay-control/
 │       ├── spec.md          - Feature specification
 │       ├── plan.md          - Implementation plan
-│       ├── tasks.md         - Task breakdown (102 tasks)
+│       ├── tasks.org        - Task breakdown (org-mode format)
-│       ├── domain-layer-architecture.md - Domain layer docs (NEW in Phase 2)
+│       ├── data-model.md    - Data model specification
-│       ├── lessons-learned.md - Phase 2 insights (NEW in Phase 2)
+│       ├── types-design.md  - Domain types design
-│       └── research-cors.md - CORS configuration research
+│       ├── domain-layer-architecture.md - Domain layer docs
+│       └── lessons-learned.md - Phase 2/3 insights
 ├── package.json             - Frontend dependencies
 ├── vite.config.ts          - Vite build configuration
 └── justfile                 - Build commands
@@ -258,17 +301,15 @@ sta/                          # Repository root
 - Rust 2024 edition
 - Poem 3.1 (web framework with OpenAPI support)
 - Tokio 1.48 (async runtime)
+- tokio-modbus (Modbus TCP client for relay hardware)
+- SQLx 0.8 (async SQLite with compile-time SQL verification)
+- async-trait (async methods in traits)
 - config (YAML configuration)
 - tracing + tracing-subscriber (structured logging)
 - governor (rate limiting)
 - thiserror (error handling)
 - serde + serde_yaml (configuration deserialization)
 
-**Planned Dependencies:**
-- tokio-modbus 0.17 (Modbus TCP client)
-- SQLx 0.8 (async SQLite database access)
-- mockall 0.13 (mocking for tests)
-
 **Frontend** (scaffolding complete):
 - Vue 3 + TypeScript
 - Vite build tool
@@ -306,6 +347,26 @@ sta/                          # Repository root
 
 **Test Coverage Achieved**: 100% domain layer coverage with comprehensive test suites
 
+**Phase 3 Infrastructure Testing:**
+- **MockRelayController Tests**: 6 tests in `mock_controller.rs`
+  - Read/write state operations
+  - Read/write all relay states
+  - Invalid relay ID handling
+  - Thread-safe concurrent access
+- **ModbusRelayController Tests**: Hardware integration tests (#[ignore])
+  - Real hardware communication tests
+  - Connection timeout handling
+- **SqliteRelayLabelRepository Tests**: Database layer tests
+  - CRUD operations on relay labels
+  - In-memory database for fast tests
+  - Compile-time SQL verification
+- **HealthMonitor Tests**: 15+ tests in `health_monitor.rs`
+  - State transitions (Healthy -> Degraded -> Unhealthy)
+  - Recovery from failure states
+  - Concurrent access safety
+
+**Test Coverage Achieved**: Comprehensive coverage across all layers with TDD approach
+
 ## Documentation
 
 ### Configuration Guides

backend/.tarpaulin.local.toml

@@ -4,4 +4,4 @@ skip-clean = true
 target-dir = "coverage"
 output-dir = "coverage"
 fail-under = 60
-exclude-files = ["target/*", "private/*", "tests/*"]
+exclude-files = ["target/*", "private/*", "backend/tests/*", "backend/build.rs"]

backend/settings/base.yaml

@@ -8,7 +8,7 @@ rate_limit:
   per_seconds: 60
 
 modbus:
-  host: "192.168.0.200"
+  host: 192.168.0.200
   port: 502
   slave_id: 0
   timeout_secs: 5

backend/src/application/health/health_monitor.rs

@@ -0,0 +1,331 @@
+//! Health monitoring service for tracking system health status.
+//!
+//! The `HealthMonitor` service tracks the health status of the Modbus relay controller
+//! by monitoring consecutive errors and transitions between healthy, degraded, and unhealthy states.
+//! This service implements the health monitoring requirements from FR-020 and FR-021.
+
+use std::sync::Arc;
+use tokio::sync::Mutex;
+
+use crate::domain::health::HealthStatus;
+
+/// Health monitor service for tracking system health status.
+///
+/// The `HealthMonitor` service maintains the current health status and provides
+/// methods to track successes and failures, transitioning between states according
+/// to the business rules defined in the domain layer.
+#[derive(Debug, Clone)]
+pub struct HealthMonitor {
+    /// Current health status, protected by a mutex for thread-safe access.
+    current_status: Arc<Mutex<HealthStatus>>,
+}
+
+impl HealthMonitor {
+    /// Creates a new `HealthMonitor` with initial `Healthy` status.
+    #[must_use]
+    pub fn new() -> Self {
+        Self::with_initial_status(HealthStatus::Healthy)
+    }
+
+    /// Creates a new `HealthMonitor` with the specified initial status.
+    #[must_use]
+    pub fn with_initial_status(initial_status: HealthStatus) -> Self {
+        Self {
+            current_status: Arc::new(Mutex::new(initial_status)),
+        }
+    }
+
+    /// Records a successful operation, potentially transitioning to `Healthy` status.
+    ///
+    /// This method transitions the health status according to the following rules:
+    /// - If currently `Healthy`: remains `Healthy`
+    /// - If currently `Degraded`: transitions to `Healthy` (recovery)
+    /// - If currently `Unhealthy`: transitions to `Healthy` (recovery)
+    ///
+    /// # Returns
+    ///
+    /// The new health status after recording the success.
+    pub async fn track_success(&self) -> HealthStatus {
+        let mut status = self.current_status.lock().await;
+        let new_status = status.clone().record_success();
+        *status = new_status.clone();
+        new_status
+    }
+
+    /// Records a failed operation, potentially transitioning to `Degraded` or `Unhealthy` status.
+    ///
+    /// This method transitions the health status according to the following rules:
+    /// - If currently `Healthy`: transitions to `Degraded` with 1 consecutive error
+    /// - If currently `Degraded`: increments consecutive error count
+    /// - If currently `Unhealthy`: remains `Unhealthy`
+    ///
+    /// # Returns
+    ///
+    /// The new health status after recording the failure.
+    pub async fn track_failure(&self) -> HealthStatus {
+        let mut status = self.current_status.lock().await;
+        let new_status = status.clone().record_error();
+        *status = new_status.clone();
+        new_status
+    }
+
+    /// Marks the system as unhealthy with the specified reason.
+    ///
+    /// This method immediately transitions to `Unhealthy` status regardless of
+    /// the current status, providing a way to explicitly mark critical failures.
+    ///
+    /// # Parameters
+    ///
+    /// - `reason`: Human-readable description of the failure reason.
+    ///
+    /// # Returns
+    ///
+    /// The new `Unhealthy` health status.
+    pub async fn mark_unhealthy(&self, reason: impl Into<String>) -> HealthStatus {
+        let mut status = self.current_status.lock().await;
+        let new_status = status.clone().mark_unhealthy(reason);
+        *status = new_status.clone();
+        new_status
+    }
+
+    /// Gets the current health status without modifying it.
+    ///
+    /// # Returns
+    ///
+    /// The current health status.
+    pub async fn get_status(&self) -> HealthStatus {
+        let status = self.current_status.lock().await;
+        status.clone()
+    }
+
+    /// Checks if the system is currently healthy.
+    ///
+    /// # Returns
+    ///
+    /// `true` if the current status is `Healthy`, `false` otherwise.
+    pub async fn is_healthy(&self) -> bool {
+        let status = self.current_status.lock().await;
+        status.is_healthy()
+    }
+
+    /// Checks if the system is currently degraded.
+    ///
+    /// # Returns
+    ///
+    /// `true` if the current status is `Degraded`, `false` otherwise.
+    pub async fn is_degraded(&self) -> bool {
+        let status = self.current_status.lock().await;
+        status.is_degraded()
+    }
+
+    /// Checks if the system is currently unhealthy.
+    ///
+    /// # Returns
+    ///
+    /// `true` if the current status is `Unhealthy`, `false` otherwise.
+    pub async fn is_unhealthy(&self) -> bool {
+        let status = self.current_status.lock().await;
+        status.is_unhealthy()
+    }
+}
+
+impl Default for HealthMonitor {
+    fn default() -> Self {
+        Self::new()
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[tokio::test]
+    async fn test_health_monitor_initial_state() {
+        let monitor = HealthMonitor::new();
+        let status = monitor.get_status().await;
+        assert!(status.is_healthy());
+    }
+
+    #[tokio::test]
+    async fn test_health_monitor_with_initial_status() {
+        let initial_status = HealthStatus::degraded(3);
+        let monitor = HealthMonitor::with_initial_status(initial_status.clone());
+        let status = monitor.get_status().await;
+        assert_eq!(status, initial_status);
+    }
+
+    #[tokio::test]
+    async fn test_track_success_from_healthy() {
+        let monitor = HealthMonitor::new();
+        let status = monitor.track_success().await;
+        assert!(status.is_healthy());
+    }
+
+    #[tokio::test]
+    async fn test_track_success_from_degraded() {
+        let monitor = HealthMonitor::with_initial_status(HealthStatus::degraded(5));
+        let status = monitor.track_success().await;
+        assert!(status.is_healthy());
+    }
+
+    #[tokio::test]
+    async fn test_track_success_from_unhealthy() {
+        let monitor = HealthMonitor::with_initial_status(HealthStatus::unhealthy("Test failure"));
+        let status = monitor.track_success().await;
+        assert!(status.is_healthy());
+    }
+
+    #[tokio::test]
+    async fn test_track_failure_from_healthy() {
+        let monitor = HealthMonitor::new();
+        let status = monitor.track_failure().await;
+        assert!(status.is_degraded());
+        assert_eq!(status, HealthStatus::degraded(1));
+    }
+
+    #[tokio::test]
+    async fn test_track_failure_from_degraded() {
+        let monitor = HealthMonitor::with_initial_status(HealthStatus::degraded(2));
+        let status = monitor.track_failure().await;
+        assert!(status.is_degraded());
+        assert_eq!(status, HealthStatus::degraded(3));
+    }
+
+    #[tokio::test]
+    async fn test_track_failure_from_unhealthy() {
+        let monitor =
+            HealthMonitor::with_initial_status(HealthStatus::unhealthy("Critical failure"));
+        let status = monitor.track_failure().await;
+        assert!(status.is_unhealthy());
+        assert_eq!(status, HealthStatus::unhealthy("Critical failure"));
+    }
+
+    #[tokio::test]
+    async fn test_mark_unhealthy() {
+        let monitor = HealthMonitor::new();
+        let status = monitor.mark_unhealthy("Device disconnected").await;
+        assert!(status.is_unhealthy());
+        assert_eq!(status, HealthStatus::unhealthy("Device disconnected"));
+    }
+
+    #[tokio::test]
+    async fn test_mark_unhealthy_overwrites_previous() {
+        let monitor = HealthMonitor::with_initial_status(HealthStatus::degraded(3));
+        let status = monitor.mark_unhealthy("New failure").await;
+        assert!(status.is_unhealthy());
+        assert_eq!(status, HealthStatus::unhealthy("New failure"));
+    }
+
+    #[tokio::test]
+    async fn test_get_status() {
+        let monitor = HealthMonitor::with_initial_status(HealthStatus::degraded(2));
+        let status = monitor.get_status().await;
+        assert_eq!(status, HealthStatus::degraded(2));
+    }
+
+    #[tokio::test]
+    async fn test_is_healthy() {
+        let healthy_monitor = HealthMonitor::new();
+        assert!(healthy_monitor.is_healthy().await);
+
+        let degraded_monitor = HealthMonitor::with_initial_status(HealthStatus::degraded(1));
+        assert!(!degraded_monitor.is_healthy().await);
+
+        let unhealthy_monitor =
+            HealthMonitor::with_initial_status(HealthStatus::unhealthy("Failure"));
+        assert!(!unhealthy_monitor.is_healthy().await);
+    }
+
+    #[tokio::test]
+    async fn test_is_degraded() {
+        let healthy_monitor = HealthMonitor::new();
+        assert!(!healthy_monitor.is_degraded().await);
+
+        let degraded_monitor = HealthMonitor::with_initial_status(HealthStatus::degraded(1));
+        assert!(degraded_monitor.is_degraded().await);
+
+        let unhealthy_monitor =
+            HealthMonitor::with_initial_status(HealthStatus::unhealthy("Failure"));
+        assert!(!unhealthy_monitor.is_degraded().await);
+    }
+
+    #[tokio::test]
+    async fn test_is_unhealthy() {
+        let healthy_monitor = HealthMonitor::new();
+        assert!(!healthy_monitor.is_unhealthy().await);
+
+        let degraded_monitor = HealthMonitor::with_initial_status(HealthStatus::degraded(1));
+        assert!(!degraded_monitor.is_unhealthy().await);
+
+        let unhealthy_monitor =
+            HealthMonitor::with_initial_status(HealthStatus::unhealthy("Failure"));
+        assert!(unhealthy_monitor.is_unhealthy().await);
+    }
+
+    #[tokio::test]
+    async fn test_state_transitions_sequence() {
+        let monitor = HealthMonitor::new();
+
+        // Start healthy
+        assert!(monitor.is_healthy().await);
+
+        // First failure -> Degraded with 1 error
+        let status = monitor.track_failure().await;
+        assert!(status.is_degraded());
+        assert_eq!(status, HealthStatus::degraded(1));
+
+        // Second failure -> Degraded with 2 errors
+        let status = monitor.track_failure().await;
+        assert_eq!(status, HealthStatus::degraded(2));
+
+        // Third failure -> Degraded with 3 errors
+        let status = monitor.track_failure().await;
+        assert_eq!(status, HealthStatus::degraded(3));
+
+        // Recovery -> Healthy
+        let status = monitor.track_success().await;
+        assert!(status.is_healthy());
+
+        // Another failure -> Degraded with 1 error
+        let status = monitor.track_failure().await;
+        assert_eq!(status, HealthStatus::degraded(1));
+
+        // Mark as unhealthy -> Unhealthy
+        let status = monitor.mark_unhealthy("Critical error").await;
+        assert!(status.is_unhealthy());
+
+        // Recovery from unhealthy -> Healthy
+        let status = monitor.track_success().await;
+        assert!(status.is_healthy());
+    }
+
+    #[tokio::test]
+    async fn test_concurrent_access() {
+        let monitor = HealthMonitor::new();
+
+        // Create multiple tasks that access the monitor concurrently
+        // We need to clone the monitor for each task since tokio::spawn requires 'static
+        let monitor1 = monitor.clone();
+        let monitor2 = monitor.clone();
+        let monitor3 = monitor.clone();
+        let monitor4 = monitor.clone();
+
+        let task1 = tokio::spawn(async move { monitor1.track_failure().await });
+        let task2 = tokio::spawn(async move { monitor2.track_failure().await });
+        let task3 = tokio::spawn(async move { monitor3.track_success().await });
+        let task4 = tokio::spawn(async move { monitor4.get_status().await });
+
+        // Wait for all tasks to complete
+        let (result1, result2, result3, result4) = tokio::join!(task1, task2, task3, task4);
+
+        // All operations should complete without panicking
+        result1.expect("Task should complete successfully");
+        result2.expect("Task should complete successfully");
+        result3.expect("Task should complete successfully");
+        result4.expect("Task should complete successfully");
+
+        // Final status should be healthy (due to the success operation)
+        let final_status = monitor.get_status().await;
+        assert!(final_status.is_healthy());
+    }
+}

backend/src/application/health/mod.rs

@@ -0,0 +1,6 @@
+//! Health monitoring application layer.
+//!
+//! This module contains the health monitoring service that tracks the system's
+//! health status and manages state transitions between healthy, degraded, and unhealthy states.
+
+pub mod health_monitor;

backend/src/application/mod.rs

@@ -11,6 +11,11 @@
 //! - **Use case driven**: Each module represents a specific business use case
 //! - **Testable in isolation**: Can be tested with mock infrastructure implementations
 //!
+//! # Submodules
+//!
+//! - `health`: Health monitoring service
+//!   - `health_monitor`: Tracks system health status and state transitions
+//!
 //! # Planned Submodules
 //!
 //! - `relay`: Relay control use cases
@@ -58,3 +63,5 @@
 //! - Architecture: `specs/constitution.md` - Hexagonal Architecture principles
 //! - Use cases: `specs/001-modbus-relay-control/plan.md` - Implementation plan
 //! - Domain types: [`crate::domain`] - Domain entities and value objects
+
+pub mod health;

backend/src/domain/relay/repository/mod.rs

@@ -1,7 +1,7 @@
 mod label;
 pub use label::RelayLabelRepository;
 
-use super::types::RelayId;
+use super::types::{RelayId, RelayLabelError};
 
 /// Errors that can occur during repository operations.
 ///
@@ -16,3 +16,15 @@ pub enum RepositoryError {
     #[error("Relay not found: {0}")]
     NotFound(RelayId),
 }
+
+impl From<sqlx::Error> for RepositoryError {
+    fn from(value: sqlx::Error) -> Self {
+        Self::DatabaseError(value.to_string())
+    }
+}
+
+impl From<RelayLabelError> for RepositoryError {
+    fn from(value: RelayLabelError) -> Self {
+        Self::DatabaseError(value.to_string())
+    }
+}

backend/src/domain/relay/types/mod.rs

@@ -3,5 +3,5 @@ mod relaylabel;
 mod relaystate;
 
 pub use relayid::RelayId;
-pub use relaylabel::RelayLabel;
+pub use relaylabel::{RelayLabel, RelayLabelError};
 pub use relaystate::RelayState;

backend/src/domain/relay/types/relaylabel.rs

@@ -8,10 +8,19 @@ use thiserror::Error;
 #[repr(transparent)]
 pub struct RelayLabel(String);
 
+/// Errors that can occur when creating or validating relay labels.
 #[derive(Debug, Error)]
 pub enum RelayLabelError {
+    /// The label string is empty.
+    ///
+    /// Relay labels must contain at least one character.
     #[error("Label cannot be empty")]
     Empty,
+
+    /// The label string exceeds the maximum allowed length.
+    ///
+    /// Contains the actual length of the invalid label.
+    /// Maximum allowed length is 50 characters.
     #[error("Label exceeds maximum length of 50 characters: {0}")]
     TooLong(usize),
 }

backend/src/infrastructure/modbus/client_test.rs

@@ -10,6 +10,10 @@ use super::*;
 mod t025a_connection_setup_tests {
     use super::*;
 
+    static HOST: &str = "192.168.1.200";
+    static PORT: u16 = 502;
+    static SLAVE_ID: u8 = 1;
+
     /// T025a Test 1: `new()` with valid config connects successfully
     ///
     /// This test verifies that `ModbusRelayController::new()` can establish
@@ -21,13 +25,10 @@ mod t025a_connection_setup_tests {
     #[ignore = "Requires running Modbus TCP server"]
     async fn test_new_with_valid_config_connects_successfully() {
         // Arrange: Use localhost test server
-        let host = "127.0.0.1";
-        let port = 5020; // Test Modbus TCP port
-        let slave_id = 1;
         let timeout_secs = 5;
 
         // Act: Attempt to create controller
-        let result = ModbusRelayController::new(host, port, slave_id, timeout_secs).await;
+        let result = ModbusRelayController::new(HOST, PORT, SLAVE_ID, timeout_secs).await;
 
         // Assert: Connection should succeed
         assert!(
@@ -45,12 +46,10 @@ mod t025a_connection_setup_tests {
     async fn test_new_with_invalid_host_returns_connection_error() {
         // Arrange: Use invalid host format
         let host = "not a valid host!!!";
-        let port = 502;
-        let slave_id = 1;
         let timeout_secs = 5;
 
         // Act: Attempt to create controller
-        let result = ModbusRelayController::new(host, port, slave_id, timeout_secs).await;
+        let result = ModbusRelayController::new(host, PORT, SLAVE_ID, timeout_secs).await;
 
         // Assert: Should return ConnectionError
         assert!(result.is_err(), "Expected ConnectionError for invalid host");
@@ -74,13 +73,11 @@ mod t025a_connection_setup_tests {
     async fn test_new_with_unreachable_host_returns_connection_error() {
         // Arrange: Use localhost with a closed port (port 1 is typically closed)
         // This gives instant "connection refused" instead of waiting for TCP timeout
-        let host = "127.0.0.1";
         let port = 1; // Closed port for instant connection failure
-        let slave_id = 1;
         let timeout_secs = 1;
 
         // Act: Attempt to create controller
-        let result = ModbusRelayController::new(host, port, slave_id, timeout_secs).await;
+        let result = ModbusRelayController::new(HOST, port, SLAVE_ID, timeout_secs).await;
 
         // Assert: Should return ConnectionError
         assert!(
@@ -100,13 +97,10 @@ mod t025a_connection_setup_tests {
     #[ignore = "Requires running Modbus TCP server or refactoring to expose timeout"]
     async fn test_new_stores_correct_timeout_duration() {
         // Arrange
-        let host = "127.0.0.1";
-        let port = 5020;
-        let slave_id = 1;
         let timeout_secs = 10;
 
         // Act
-        let controller = ModbusRelayController::new(host, port, slave_id, timeout_secs)
+        let controller = ModbusRelayController::new(HOST, PORT, SLAVE_ID, timeout_secs)
             .await
             .expect("Failed to create controller");
 
@@ -137,6 +131,10 @@ mod t025b_read_coils_timeout_tests {
         types::RelayId,
     };
 
+    static HOST: &str = "192.168.1.200";
+    static PORT: u16 = 502;
+    static SLAVE_ID: u8 = 1;
+
     /// T025b Test 1: `read_coils_with_timeout()` returns coil values on success
     ///
     /// This test verifies that reading coils succeeds when the Modbus server
@@ -147,7 +145,7 @@ mod t025b_read_coils_timeout_tests {
     #[ignore = "Requires running Modbus TCP server with known state"]
     async fn test_read_coils_returns_coil_values_on_success() {
         // Arrange: Connect to test server
-        let controller = ModbusRelayController::new("127.0.0.1", 5020, 1, 5)
+        let controller = ModbusRelayController::new(HOST, PORT, SLAVE_ID, 5)
             .await
             .expect("Failed to connect to test server");
 
@@ -251,6 +249,10 @@ mod t025c_write_single_coil_timeout_tests {
         types::{RelayId, RelayState},
     };
 
+    static HOST: &str = "192.168.1.200";
+    static PORT: u16 = 502;
+    static SLAVE_ID: u8 = 1;
+
     /// T025c Test 1: `write_single_coil_with_timeout()` succeeds for valid write
     ///
     /// This test verifies that writing to a coil succeeds when the Modbus server
@@ -261,7 +263,7 @@ mod t025c_write_single_coil_timeout_tests {
     #[ignore = "Requires running Modbus TCP server"]
     async fn test_write_single_coil_succeeds_for_valid_write() {
         // Arrange: Connect to test server
-        let controller = ModbusRelayController::new("127.0.0.1", 5020, 1, 5)
+        let controller = ModbusRelayController::new(HOST, PORT, SLAVE_ID, 5)
             .await
             .expect("Failed to connect to test server");
 
@@ -336,6 +338,10 @@ mod t025d_read_relay_state_tests {
         types::{RelayId, RelayState},
     };
 
+    static HOST: &str = "192.168.1.200";
+    static PORT: u16 = 502;
+    static SLAVE_ID: u8 = 1;
+
     /// T025d Test 1: `read_relay_state(RelayId(1))` returns On when coil is true
     ///
     /// This test verifies that a true coil value is correctly converted to `RelayState::On`.
@@ -409,7 +415,7 @@ mod t025d_read_relay_state_tests {
     #[ignore = "Requires Modbus server with specific relay states"]
     async fn test_read_state_correctly_maps_relay_id_to_modbus_address() {
         // Arrange: Connect to test server with known relay states
-        let controller = ModbusRelayController::new("127.0.0.1", 5020, 1, 5)
+        let controller = ModbusRelayController::new(HOST, PORT, SLAVE_ID, 5)
             .await
             .expect("Failed to connect to test server");
 
@@ -434,6 +440,10 @@ mod t025e_write_relay_state_tests {
         types::{RelayId, RelayState},
     };
 
+    static HOST: &str = "192.168.1.200";
+    static PORT: u16 = 502;
+    static SLAVE_ID: u8 = 1;
+
     /// T025e Test 1: `write_relay_state(RelayId(1)`, `RelayState::On`) writes true to coil
     ///
     /// This test verifies that `RelayState::On` is correctly converted to a true coil value.
@@ -441,7 +451,7 @@ mod t025e_write_relay_state_tests {
     #[ignore = "Requires Modbus server that can verify written values"]
     async fn test_write_state_on_writes_true_to_coil() {
         // Arrange: Connect to test server
-        let controller = ModbusRelayController::new("127.0.0.1", 5020, 1, 5)
+        let controller = ModbusRelayController::new(HOST, PORT, SLAVE_ID, 5)
             .await
             .expect("Failed to connect to test server");
 
@@ -475,7 +485,7 @@ mod t025e_write_relay_state_tests {
     #[ignore = "Requires Modbus server that can verify written values"]
     async fn test_write_state_off_writes_false_to_coil() {
         // Arrange: Connect to test server
-        let controller = ModbusRelayController::new("127.0.0.1", 5020, 1, 5)
+        let controller = ModbusRelayController::new(HOST, PORT, SLAVE_ID, 5)
             .await
             .expect("Failed to connect to test server");
 
@@ -509,7 +519,7 @@ mod t025e_write_relay_state_tests {
     #[ignore = "Requires Modbus server"]
     async fn test_write_state_correctly_maps_relay_id_to_modbus_address() {
         // Arrange: Connect to test server
-        let controller = ModbusRelayController::new("127.0.0.1", 5020, 1, 5)
+        let controller = ModbusRelayController::new(HOST, PORT, SLAVE_ID, 5)
             .await
             .expect("Failed to connect to test server");
 
@@ -537,7 +547,7 @@ mod t025e_write_relay_state_tests {
     #[ignore = "Requires Modbus server"]
     async fn test_write_state_can_toggle_relay_multiple_times() {
         // Arrange: Connect to test server
-        let controller = ModbusRelayController::new("127.0.0.1", 5020, 1, 5)
+        let controller = ModbusRelayController::new(HOST, PORT, SLAVE_ID, 5)
             .await
             .expect("Failed to connect to test server");
 
@@ -571,12 +581,16 @@ mod t025e_write_relay_state_tests {
 mod write_all_states_validation_tests {
     use super::*;
 
+    static HOST: &str = "192.168.1.200";
+    static PORT: u16 = 502;
+    static SLAVE_ID: u8 = 1;
+
     /// Test: `write_all_states()` returns `InvalidInput` when given 0 states
     #[tokio::test]
     #[ignore = "Requires Modbus server"]
     async fn test_write_all_states_with_empty_vector_returns_invalid_input() {
         // Arrange: Connect to test server
-        let controller = ModbusRelayController::new("127.0.0.1", 5020, 1, 5)
+        let controller = ModbusRelayController::new(HOST, PORT, SLAVE_ID, 5)
             .await
             .expect("Failed to connect to test server");
 
@@ -596,7 +610,7 @@ mod write_all_states_validation_tests {
     #[ignore = "Requires Modbus server"]
     async fn test_write_all_states_with_7_states_returns_invalid_input() {
         // Arrange: Connect to test server
-        let controller = ModbusRelayController::new("127.0.0.1", 5020, 1, 5)
+        let controller = ModbusRelayController::new(HOST, PORT, SLAVE_ID, 5)
             .await
             .expect("Failed to connect to test server");
 
@@ -626,7 +640,7 @@ mod write_all_states_validation_tests {
     #[ignore = "Requires Modbus server"]
     async fn test_write_all_states_with_9_states_returns_invalid_input() {
         // Arrange: Connect to test server
-        let controller = ModbusRelayController::new("127.0.0.1", 5020, 1, 5)
+        let controller = ModbusRelayController::new(HOST, PORT, SLAVE_ID, 5)
             .await
             .expect("Failed to connect to test server");
 
@@ -656,7 +670,7 @@ mod write_all_states_validation_tests {
     #[ignore = "Requires Modbus server"]
     async fn test_write_all_states_with_8_states_succeeds() {
         // Arrange: Connect to test server
-        let controller = ModbusRelayController::new("127.0.0.1", 5020, 1, 5)
+        let controller = ModbusRelayController::new(HOST, PORT, SLAVE_ID, 5)
             .await
             .expect("Failed to connect to test server");
 

backend/src/infrastructure/persistence/entities/mod.rs

@@ -0,0 +1,33 @@
+//! Infrastructure entities for database persistence.
+//!
+//! This module defines entities that directly map to database tables,
+//! providing a clear separation between the persistence layer and the
+//! domain layer. These entities represent raw database records without
+//! domain validation or business logic.
+//!
+//! # Conversion Pattern
+//!
+//! Infrastructure entities implement `TryFrom` traits to convert between
+//! database records and domain types:
+//!
+//! ```rust
+//! # use sta::domain::relay::types::{RelayId, RelayLabel};
+//! # use sta::infrastructure::persistence::entities::relay_label_record::RelayLabelRecord;
+//! # fn main() -> Result<(), Box<dyn std::error::Error>> {
+//! // Database Record -> Domain Types
+//! // ... from database
+//! let record: RelayLabelRecord = RelayLabelRecord { relay_id: 2, label: "label".to_string() };
+//! let (relay_id, relay_label): (RelayId, RelayLabel) = record.try_into()?;
+//!
+//! // Domain Types -> Database Record
+//! let domain_record= RelayLabelRecord::new(relay_id, &relay_label);
+//! # Ok(())
+//! # }
+//! ```
+
+/// Database entity for relay labels.
+///
+/// This module contains the `RelayLabelRecord` struct which represents
+/// a single row in the `RelayLabels` database table, along with conversion
+/// traits to and from domain types.
+pub mod relay_label_record;

backend/src/infrastructure/persistence/entities/relay_label_record.rs

@@ -0,0 +1,62 @@
+use crate::domain::relay::{
+    controller::ControllerError,
+    repository::RepositoryError,
+    types::{RelayId, RelayLabel, RelayLabelError},
+};
+
+/// Database record representing a relay label.
+///
+/// This struct directly maps to the `RelayLabels` table in the
+/// database. It represents the raw data as stored in the database,
+/// without domain validation or business logic.
+#[derive(Debug, Clone, sqlx::FromRow)]
+pub struct RelayLabelRecord {
+    /// The relay ID (1-8) as stored in the database
+    pub relay_id: i64,
+    /// The label text as stored in the database
+    pub label: String,
+}
+
+impl RelayLabelRecord {
+    /// Creates a new `RecordLabelRecord` from domain types.
+    #[must_use]
+    pub fn new(relay_id: RelayId, label: &RelayLabel) -> Self {
+        Self {
+            relay_id: i64::from(relay_id.as_u8()),
+            label: label.as_str().to_string(),
+        }
+    }
+}
+
+impl TryFrom<RelayLabelRecord> for RelayId {
+    type Error = ControllerError;
+
+    fn try_from(value: RelayLabelRecord) -> Result<Self, Self::Error> {
+        let value = u8::try_from(value.relay_id).map_err(|e| {
+            Self::Error::InvalidInput(format!("Got value {} from database: {e}", value.relay_id))
+        })?;
+        Self::new(value)
+    }
+}
+
+impl TryFrom<RelayLabelRecord> for RelayLabel {
+    type Error = RelayLabelError;
+
+    fn try_from(value: RelayLabelRecord) -> Result<Self, Self::Error> {
+        Self::new(value.label)
+    }
+}
+
+impl TryFrom<RelayLabelRecord> for (RelayId, RelayLabel) {
+    type Error = RepositoryError;
+
+    fn try_from(value: RelayLabelRecord) -> Result<Self, Self::Error> {
+        let record_id: RelayId = value
+            .clone()
+            .try_into()
+            .map_err(|e: ControllerError| RepositoryError::DatabaseError(e.to_string()))?;
+        let label: RelayLabel = RelayLabel::new(value.label)
+            .map_err(|e| RepositoryError::DatabaseError(e.to_string()))?;
+        Ok((record_id, label))
+    }
+}

backend/src/infrastructure/persistence/label_repository_tests.rs

@@ -124,7 +124,10 @@ mod relay_label_repository_contract_tests {
             .expect("Second save should succeed");
 
         // Verify only the second label is present
-        let result = repo.get_label(relay_id).await.expect("get_label should succeed");
+        let result = repo
+            .get_label(relay_id)
+            .await
+            .expect("get_label should succeed");
         assert!(result.is_some(), "Label should exist");
         assert_eq!(
             result.unwrap().as_str(),
@@ -270,11 +273,17 @@ mod relay_label_repository_contract_tests {
             .expect("delete should succeed");
 
         // Verify deleted label is gone
-        let get_result = repo.get_label(relay2).await.expect("get_label should succeed");
+        let get_result = repo
+            .get_label(relay2)
+            .await
+            .expect("get_label should succeed");
         assert!(get_result.is_none(), "Deleted label should not exist");
 
         // Verify other label still exists
-        let other_result = repo.get_label(relay1).await.expect("get_label should succeed");
+        let other_result = repo
+            .get_label(relay1)
+            .await
+            .expect("get_label should succeed");
         assert!(other_result.is_some(), "Other label should still exist");
 
         // Verify get_all_labels only returns the remaining label

backend/src/infrastructure/persistence/mod.rs

@@ -12,3 +12,5 @@ pub mod label_repository_tests;
 
 /// `SQLite` repository implementation for relay labels.
 pub mod sqlite_repository;
+
+pub mod entities;

backend/src/infrastructure/persistence/sqlite_repository.rs

@@ -1,6 +1,13 @@
-use sqlx::SqlitePool;
+use async_trait::async_trait;
+use sqlx::{SqlitePool, query_as};
 
-use crate::domain::relay::repository::RepositoryError;
+use crate::{
+    domain::relay::{
+        repository::{RelayLabelRepository, RepositoryError},
+        types::{RelayId, RelayLabel},
+    },
+    infrastructure::persistence::entities::relay_label_record::RelayLabelRecord,
+};
 
 /// `SQLite` implementation of the relay label repository.
 ///
@@ -62,3 +69,56 @@ impl SqliteRelayLabelRepository {
         Ok(())
     }
 }
+
+#[async_trait]
+impl RelayLabelRepository for SqliteRelayLabelRepository {
+    async fn get_label(&self, id: RelayId) -> Result<Option<RelayLabel>, RepositoryError> {
+        let id = i64::from(id.as_u8());
+        let result = sqlx::query_as!(
+            RelayLabelRecord,
+            "SELECT * FROM RelayLabels WHERE relay_id = ?1",
+            id
+        )
+        .fetch_optional(&self.pool)
+        .await
+        .map_err(|e| RepositoryError::DatabaseError(e.to_string()))?;
+
+        match result {
+            Some(record) => Ok(Some(record.try_into()?)),
+            None => Ok(None),
+        }
+    }
+
+    async fn save_label(&self, id: RelayId, label: RelayLabel) -> Result<(), RepositoryError> {
+        let record = RelayLabelRecord::new(id, &label);
+        sqlx::query!(
+            "INSERT OR REPLACE INTO RelayLabels (relay_id, label) VALUES (?1, ?2)",
+            record.relay_id,
+            record.label
+        )
+        .execute(&self.pool)
+        .await
+        .map_err(RepositoryError::from)?;
+        Ok(())
+    }
+
+    async fn delete_label(&self, id: RelayId) -> Result<(), RepositoryError> {
+        let id = i64::from(id.as_u8());
+        sqlx::query!("DELETE FROM RelayLabels WHERE relay_id = ?1", id)
+            .execute(&self.pool)
+            .await
+            .map_err(RepositoryError::from)?;
+        Ok(())
+    }
+
+    async fn get_all_labels(&self) -> Result<Vec<(RelayId, RelayLabel)>, RepositoryError> {
+        let result: Vec<RelayLabelRecord> = query_as!(
+            RelayLabelRecord,
+            "SELECT * FROM RelayLabels ORDER BY relay_id"
+        )
+        .fetch_all(&self.pool)
+        .await
+        .map_err(RepositoryError::from)?;
+        result.iter().map(|r| r.clone().try_into()).collect()
+    }
+}

backend/tests/modbus_hardware_test.rs

@@ -0,0 +1,253 @@
+// Integration tests for Modbus hardware
+// These tests require physical Modbus relay device to be connected
+// Run with: cargo test -- --ignored
+
+use std::time::Duration;
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use sta::domain::relay::controller::RelayController;
+    use sta::domain::relay::types::{RelayId, RelayState};
+    use sta::infrastructure::modbus::client::ModbusRelayController;
+
+    static HOST: &str = "192.168.1.200";
+    static PORT: u16 = 502;
+    static SLAVE_ID: u8 = 1;
+
+    #[tokio::test]
+    #[ignore = "Requires physical Modbus device"]
+    async fn test_modbus_connection() {
+        // This test verifies we can connect to the actual Modbus device
+        // Configured with settings from settings/base.yaml
+        let timeout_secs = 5;
+
+        let _controller = ModbusRelayController::new(HOST, PORT, SLAVE_ID, timeout_secs)
+            .await
+            .expect("Failed to connect to Modbus device");
+
+        // If we got here, connection was successful
+        println!("✓ Successfully connected to Modbus device");
+    }
+
+    #[tokio::test]
+    #[ignore = "Requires physical Modbus device"]
+    async fn test_read_relay_states() {
+        let timeout_secs = 5;
+
+        let controller = ModbusRelayController::new(HOST, PORT, SLAVE_ID, timeout_secs)
+            .await
+            .expect("Failed to connect to Modbus device");
+
+        // Test reading individual relay states
+        for relay_id in 1..=8 {
+            let relay_id = RelayId::new(relay_id).unwrap();
+            let state = controller
+                .read_relay_state(relay_id)
+                .await
+                .expect("Failed to read relay state");
+
+            println!("Relay {}: {:?}", relay_id.as_u8(), state);
+        }
+    }
+
+    #[tokio::test]
+    #[ignore = "Requires physical Modbus device"]
+    async fn test_read_all_relays() {
+        let timeout_secs = 5;
+
+        let controller = ModbusRelayController::new(HOST, PORT, SLAVE_ID, timeout_secs)
+            .await
+            .expect("Failed to connect to Modbus device");
+
+        let relays = controller
+            .read_all_states()
+            .await
+            .expect("Failed to read all relay states");
+
+        assert_eq!(relays.len(), 8, "Should have exactly 8 relays");
+
+        for (i, state) in relays.iter().enumerate() {
+            let relay_id = i + 1;
+            println!("Relay {}: {:?}", relay_id, state);
+        }
+    }
+
+    #[tokio::test]
+    #[ignore = "Requires physical Modbus device"]
+    async fn test_write_relay_state() {
+        let timeout_secs = 5;
+
+        let controller = ModbusRelayController::new(HOST, PORT, SLAVE_ID, timeout_secs)
+            .await
+            .expect("Failed to connect to Modbus device");
+
+        let relay_id = RelayId::new(1).unwrap();
+
+        // Turn relay on
+        controller
+            .write_relay_state(relay_id, RelayState::On)
+            .await
+            .expect("Failed to write relay state");
+
+        // Verify it's on
+        let state = controller
+            .read_relay_state(relay_id)
+            .await
+            .expect("Failed to read relay state");
+
+        assert_eq!(state, RelayState::On, "Relay should be ON");
+
+        // Turn relay off
+        controller
+            .write_relay_state(relay_id, RelayState::Off)
+            .await
+            .expect("Failed to write relay state");
+
+        // Verify it's off
+        let state = controller
+            .read_relay_state(relay_id)
+            .await
+            .expect("Failed to read relay state");
+
+        assert_eq!(state, RelayState::Off, "Relay should be OFF");
+    }
+
+    #[tokio::test]
+    #[ignore = "Requires physical Modbus device"]
+    async fn test_write_all_relays() {
+        let timeout_secs = 5;
+
+        let controller = ModbusRelayController::new(HOST, PORT, SLAVE_ID, timeout_secs)
+            .await
+            .expect("Failed to connect to Modbus device");
+
+        // Turn all relays on
+        let all_on_states = vec![RelayState::On; 8];
+        controller
+            .write_all_states(all_on_states)
+            .await
+            .expect("Failed to write all relay states");
+
+        // Verify all are on
+        let relays = controller
+            .read_all_states()
+            .await
+            .expect("Failed to read all relay states");
+
+        for state in &relays {
+            assert_eq!(*state, RelayState::On, "All relays should be ON");
+        }
+
+        // Turn all relays off
+        let all_off_states = vec![RelayState::Off; 8];
+        controller
+            .write_all_states(all_off_states)
+            .await
+            .expect("Failed to write all relay states");
+
+        // Verify all are off
+        let relays = controller
+            .read_all_states()
+            .await
+            .expect("Failed to read all relay states");
+
+        for state in &relays {
+            assert_eq!(*state, RelayState::Off, "All relays should be OFF");
+        }
+    }
+
+    #[tokio::test]
+    #[ignore = "Requires physical Modbus device"]
+    async fn test_timeout_handling() {
+        let timeout_secs = 1; // Short timeout for testing
+
+        let controller = ModbusRelayController::new(HOST, PORT, SLAVE_ID, timeout_secs)
+            .await
+            .expect("Failed to connect to Modbus device");
+
+        // This test verifies that timeout works correctly
+        // We'll try to read a relay state with a very short timeout
+        let relay_id = RelayId::new(1).unwrap();
+
+        // The operation should either succeed quickly or timeout
+        let result = tokio::time::timeout(
+            Duration::from_secs(2),
+            controller.read_relay_state(relay_id),
+        )
+        .await;
+
+        match result {
+            Ok(Ok(state)) => {
+                println!("✓ Operation completed within timeout: {:?}", state);
+            }
+            Ok(Err(e)) => {
+                println!("✓ Operation failed (expected): {}", e);
+            }
+            Err(_) => {
+                println!("✓ Operation timed out (expected)");
+            }
+        }
+    }
+
+    #[tokio::test]
+    #[ignore = "Requires physical Modbus device"]
+    async fn test_concurrent_access() {
+        let timeout_secs = 5;
+
+        let _controller = ModbusRelayController::new(HOST, PORT, SLAVE_ID, timeout_secs)
+            .await
+            .expect("Failed to connect to Modbus device");
+
+        // Test concurrent access to the controller
+        // We'll test a few relays concurrently using tokio::join!
+        // Note: We can't clone the controller, so we'll just test sequential access
+        // This is still valuable for testing the controller works with multiple relays
+
+        let relay_id1 = RelayId::new(1).unwrap();
+        let relay_id2 = RelayId::new(2).unwrap();
+        let relay_id3 = RelayId::new(3).unwrap();
+        let relay_id4 = RelayId::new(4).unwrap();
+
+        let task1 = tokio::spawn(async move {
+            let controller = ModbusRelayController::new(HOST, PORT, SLAVE_ID, timeout_secs)
+                .await
+                .expect("Failed to connect");
+            controller.read_relay_state(relay_id1).await
+        });
+        let task2 = tokio::spawn(async move {
+            let controller = ModbusRelayController::new(HOST, PORT, SLAVE_ID, timeout_secs)
+                .await
+                .expect("Failed to connect");
+            controller.read_relay_state(relay_id2).await
+        });
+        let task3 = tokio::spawn(async move {
+            let controller = ModbusRelayController::new(HOST, PORT, SLAVE_ID, timeout_secs)
+                .await
+                .expect("Failed to connect");
+            controller.read_relay_state(relay_id3).await
+        });
+        let task4 = tokio::spawn(async move {
+            let controller = ModbusRelayController::new(HOST, PORT, SLAVE_ID, timeout_secs)
+                .await
+                .expect("Failed to connect");
+            controller.read_relay_state(relay_id4).await
+        });
+
+        let (result1, result2, result3, result4) = tokio::join!(task1, task2, task3, task4);
+
+        // Process results
+        if let Ok(Ok(state)) = result1 {
+            println!("Relay 1: {:?}", state);
+        }
+        if let Ok(Ok(state)) = result2 {
+            println!("Relay 2: {:?}", state);
+        }
+        if let Ok(Ok(state)) = result3 {
+            println!("Relay 3: {:?}", state);
+        }
+        if let Ok(Ok(state)) = result4 {
+            println!("Relay 4: {:?}", state);
+        }
+    }
+}

backend/tests/sqlite_repository_functional_test.rs

@@ -0,0 +1,473 @@
+//! Functional tests for `SqliteRelayLabelRepository` implementation.
+//!
+//! These tests verify that the SQLite repository correctly implements
+//! the `RelayLabelRepository` trait using the new infrastructure entities
+//! and conversion patterns.
+
+use sta::{
+    domain::relay::{
+        repository::RelayLabelRepository,
+        types::{RelayId, RelayLabel},
+    },
+    infrastructure::persistence::{
+        entities::relay_label_record::RelayLabelRecord,
+        sqlite_repository::SqliteRelayLabelRepository,
+    },
+};
+
+/// Test that `get_label` returns None for non-existent relay.
+#[tokio::test]
+async fn test_get_label_returns_none_for_non_existent_relay() {
+    let repo = SqliteRelayLabelRepository::in_memory()
+        .await
+        .expect("Failed to create repository");
+
+    let relay_id = RelayId::new(1).expect("Valid relay ID");
+    let result = repo.get_label(relay_id).await;
+
+    assert!(result.is_ok(), "get_label should succeed");
+    assert!(
+        result.unwrap().is_none(),
+        "get_label should return None for non-existent relay"
+    );
+}
+
+/// Test that `get_label` retrieves previously saved label.
+#[tokio::test]
+async fn test_get_label_retrieves_saved_label() {
+    let repo = SqliteRelayLabelRepository::in_memory()
+        .await
+        .expect("Failed to create repository");
+
+    let relay_id = RelayId::new(2).expect("Valid relay ID");
+    let label = RelayLabel::new("Heater".to_string()).expect("Valid label");
+
+    // Save the label
+    repo.save_label(relay_id, label.clone())
+        .await
+        .expect("save_label should succeed");
+
+    // Retrieve the label
+    let result = repo.get_label(relay_id).await;
+
+    assert!(result.is_ok(), "get_label should succeed");
+    let retrieved = result.unwrap();
+    assert!(retrieved.is_some(), "get_label should return Some");
+    assert_eq!(
+        retrieved.unwrap().as_str(),
+        "Heater",
+        "Retrieved label should match saved label"
+    );
+}
+
+/// Test that `save_label` successfully saves a label.
+#[tokio::test]
+async fn test_save_label_succeeds() {
+    let repo = SqliteRelayLabelRepository::in_memory()
+        .await
+        .expect("Failed to create repository");
+
+    let relay_id = RelayId::new(1).expect("Valid relay ID");
+    let label = RelayLabel::new("Pump".to_string()).expect("Valid label");
+
+    let result = repo.save_label(relay_id, label).await;
+
+    assert!(result.is_ok(), "save_label should succeed");
+}
+
+/// Test that `save_label` overwrites existing label.
+#[tokio::test]
+async fn test_save_label_overwrites_existing_label() {
+    let repo = SqliteRelayLabelRepository::in_memory()
+        .await
+        .expect("Failed to create repository");
+
+    let relay_id = RelayId::new(4).expect("Valid relay ID");
+    let label1 = RelayLabel::new("First".to_string()).expect("Valid label");
+    let label2 = RelayLabel::new("Second".to_string()).expect("Valid label");
+
+    // Save first label
+    repo.save_label(relay_id, label1)
+        .await
+        .expect("First save should succeed");
+
+    // Overwrite with second label
+    repo.save_label(relay_id, label2)
+        .await
+        .expect("Second save should succeed");
+
+    // Verify only the second label is present
+    let result = repo
+        .get_label(relay_id)
+        .await
+        .expect("get_label should succeed");
+    assert!(result.is_some(), "Label should exist");
+    assert_eq!(
+        result.unwrap().as_str(),
+        "Second",
+        "Label should be updated to second value"
+    );
+}
+
+/// Test that `save_label` works for all valid relay IDs (1-8).
+#[tokio::test]
+async fn test_save_label_for_all_valid_relay_ids() {
+    let repo = SqliteRelayLabelRepository::in_memory()
+        .await
+        .expect("Failed to create repository");
+
+    for id in 1..=8 {
+        let relay_id = RelayId::new(id).expect("Valid relay ID");
+        let label = RelayLabel::new(format!("Relay {}", id)).expect("Valid label");
+
+        let result = repo.save_label(relay_id, label).await;
+        assert!(
+            result.is_ok(),
+            "save_label should succeed for relay ID {}",
+            id
+        );
+    }
+
+    // Verify all labels were saved
+    let all_labels = repo
+        .get_all_labels()
+        .await
+        .expect("get_all_labels should succeed");
+    assert_eq!(all_labels.len(), 8, "Should have all 8 relay labels");
+}
+
+/// Test that `save_label` accepts maximum length labels.
+#[tokio::test]
+async fn test_save_label_accepts_max_length_labels() {
+    let repo = SqliteRelayLabelRepository::in_memory()
+        .await
+        .expect("Failed to create repository");
+
+    let relay_id = RelayId::new(5).expect("Valid relay ID");
+    let max_label = RelayLabel::new("A".repeat(50)).expect("Valid max-length label");
+
+    let result = repo.save_label(relay_id, max_label).await;
+    assert!(
+        result.is_ok(),
+        "save_label should succeed with max-length label"
+    );
+
+    // Verify it was saved correctly
+    let retrieved = repo
+        .get_label(relay_id)
+        .await
+        .expect("get_label should succeed");
+    assert!(retrieved.is_some(), "Label should be saved");
+    assert_eq!(
+        retrieved.unwrap().as_str().len(),
+        50,
+        "Label should have correct length"
+    );
+}
+
+/// Test that `delete_label` succeeds for existing label.
+#[tokio::test]
+async fn test_delete_label_succeeds_for_existing_label() {
+    let repo = SqliteRelayLabelRepository::in_memory()
+        .await
+        .expect("Failed to create repository");
+
+    let relay_id = RelayId::new(7).expect("Valid relay ID");
+    let label = RelayLabel::new("ToDelete".to_string()).expect("Valid label");
+
+    // Save the label first
+    repo.save_label(relay_id, label)
+        .await
+        .expect("save_label should succeed");
+
+    // Delete it
+    let result = repo.delete_label(relay_id).await;
+    assert!(result.is_ok(), "delete_label should succeed");
+}
+
+/// Test that `delete_label` succeeds for non-existent label.
+#[tokio::test]
+async fn test_delete_label_succeeds_for_non_existent_label() {
+    let repo = SqliteRelayLabelRepository::in_memory()
+        .await
+        .expect("Failed to create repository");
+
+    let relay_id = RelayId::new(8).expect("Valid relay ID");
+
+    // Delete without saving first
+    let result = repo.delete_label(relay_id).await;
+    assert!(
+        result.is_ok(),
+        "delete_label should succeed even if label doesn't exist"
+    );
+}
+
+/// Test that `delete_label` removes label from repository.
+#[tokio::test]
+async fn test_delete_label_removes_label_from_repository() {
+    let repo = SqliteRelayLabelRepository::in_memory()
+        .await
+        .expect("Failed to create repository");
+
+    let relay1 = RelayId::new(1).expect("Valid relay ID");
+    let relay2 = RelayId::new(2).expect("Valid relay ID");
+    let label1 = RelayLabel::new("Keep".to_string()).expect("Valid label");
+    let label2 = RelayLabel::new("Remove".to_string()).expect("Valid label");
+
+    // Save two labels
+    repo.save_label(relay1, label1)
+        .await
+        .expect("save should succeed");
+    repo.save_label(relay2, label2)
+        .await
+        .expect("save should succeed");
+
+    // Delete one label
+    repo.delete_label(relay2)
+        .await
+        .expect("delete should succeed");
+
+    // Verify deleted label is gone
+    let get_result = repo
+        .get_label(relay2)
+        .await
+        .expect("get_label should succeed");
+    assert!(get_result.is_none(), "Deleted label should not exist");
+
+    // Verify other label still exists
+    let other_result = repo
+        .get_label(relay1)
+        .await
+        .expect("get_label should succeed");
+    assert!(other_result.is_some(), "Other label should still exist");
+
+    // Verify get_all_labels only returns the remaining label
+    let all_labels = repo
+        .get_all_labels()
+        .await
+        .expect("get_all_labels should succeed");
+    assert_eq!(all_labels.len(), 1, "Should only have one label remaining");
+    assert_eq!(all_labels[0].0.as_u8(), 1, "Should be relay 1");
+}
+
+/// Test that `get_all_labels` returns empty vector when no labels exist.
+#[tokio::test]
+async fn test_get_all_labels_returns_empty_when_no_labels() {
+    let repo = SqliteRelayLabelRepository::in_memory()
+        .await
+        .expect("Failed to create repository");
+
+    let result = repo.get_all_labels().await;
+
+    assert!(result.is_ok(), "get_all_labels should succeed");
+    assert!(
+        result.unwrap().is_empty(),
+        "get_all_labels should return empty vector"
+    );
+}
+
+/// Test that `get_all_labels` returns all saved labels.
+#[tokio::test]
+async fn test_get_all_labels_returns_all_saved_labels() {
+    let repo = SqliteRelayLabelRepository::in_memory()
+        .await
+        .expect("Failed to create repository");
+
+    let relay1 = RelayId::new(1).expect("Valid relay ID");
+    let relay3 = RelayId::new(3).expect("Valid relay ID");
+    let relay5 = RelayId::new(5).expect("Valid relay ID");
+
+    let label1 = RelayLabel::new("Pump".to_string()).expect("Valid label");
+    let label3 = RelayLabel::new("Heater".to_string()).expect("Valid label");
+    let label5 = RelayLabel::new("Fan".to_string()).expect("Valid label");
+
+    // Save labels
+    repo.save_label(relay1, label1.clone())
+        .await
+        .expect("Save should succeed");
+    repo.save_label(relay3, label3.clone())
+        .await
+        .expect("Save should succeed");
+    repo.save_label(relay5, label5.clone())
+        .await
+        .expect("Save should succeed");
+
+    // Retrieve all labels
+    let result = repo
+        .get_all_labels()
+        .await
+        .expect("get_all_labels should succeed");
+
+    assert_eq!(result.len(), 3, "Should return exactly 3 labels");
+
+    // Verify the labels are present (order may vary by implementation)
+    let has_relay1 = result
+        .iter()
+        .any(|(id, label)| id.as_u8() == 1 && label.as_str() == "Pump");
+    let has_relay3 = result
+        .iter()
+        .any(|(id, label)| id.as_u8() == 3 && label.as_str() == "Heater");
+    let has_relay5 = result
+        .iter()
+        .any(|(id, label)| id.as_u8() == 5 && label.as_str() == "Fan");
+
+    assert!(has_relay1, "Should contain relay 1 with label 'Pump'");
+    assert!(has_relay3, "Should contain relay 3 with label 'Heater'");
+    assert!(has_relay5, "Should contain relay 5 with label 'Fan'");
+}
+
+/// Test that `get_all_labels` excludes relays without labels.
+#[tokio::test]
+async fn test_get_all_labels_excludes_relays_without_labels() {
+    let repo = SqliteRelayLabelRepository::in_memory()
+        .await
+        .expect("Failed to create repository");
+
+    let relay2 = RelayId::new(2).expect("Valid relay ID");
+    let label2 = RelayLabel::new("Only This One".to_string()).expect("Valid label");
+
+    repo.save_label(relay2, label2)
+        .await
+        .expect("Save should succeed");
+
+    let result = repo
+        .get_all_labels()
+        .await
+        .expect("get_all_labels should succeed");
+
+    assert_eq!(
+        result.len(),
+        1,
+        "Should return only the one relay with a label"
+    );
+    assert_eq!(result[0].0.as_u8(), 2, "Should be relay 2");
+}
+
+/// Test that `get_all_labels` excludes deleted labels.
+#[tokio::test]
+async fn test_get_all_labels_excludes_deleted_labels() {
+    let repo = SqliteRelayLabelRepository::in_memory()
+        .await
+        .expect("Failed to create repository");
+
+    let relay1 = RelayId::new(1).expect("Valid relay ID");
+    let relay2 = RelayId::new(2).expect("Valid relay ID");
+    let relay3 = RelayId::new(3).expect("Valid relay ID");
+
+    let label1 = RelayLabel::new("Keep1".to_string()).expect("Valid label");
+    let label2 = RelayLabel::new("Delete".to_string()).expect("Valid label");
+    let label3 = RelayLabel::new("Keep2".to_string()).expect("Valid label");
+
+    // Save all three labels
+    repo.save_label(relay1, label1)
+        .await
+        .expect("save should succeed");
+    repo.save_label(relay2, label2)
+        .await
+        .expect("save should succeed");
+    repo.save_label(relay3, label3)
+        .await
+        .expect("save should succeed");
+
+    // Delete the middle one
+    repo.delete_label(relay2)
+        .await
+        .expect("delete should succeed");
+
+    // Verify get_all_labels only returns the two remaining labels
+    let result = repo
+        .get_all_labels()
+        .await
+        .expect("get_all_labels should succeed");
+    assert_eq!(result.len(), 2, "Should have 2 labels after deletion");
+
+    let has_relay1 = result.iter().any(|(id, _)| id.as_u8() == 1);
+    let has_relay2 = result.iter().any(|(id, _)| id.as_u8() == 2);
+    let has_relay3 = result.iter().any(|(id, _)| id.as_u8() == 3);
+
+    assert!(has_relay1, "Relay 1 should be present");
+    assert!(!has_relay2, "Relay 2 should NOT be present (deleted)");
+    assert!(has_relay3, "Relay 3 should be present");
+}
+
+/// Test that entity conversion works correctly.
+#[tokio::test]
+async fn test_entity_conversion_roundtrip() {
+    let repo = SqliteRelayLabelRepository::in_memory()
+        .await
+        .expect("Failed to create repository");
+
+    let relay_id = RelayId::new(1).expect("Valid relay ID");
+    let relay_label = RelayLabel::new("Test Label".to_string()).expect("Valid label");
+
+    // Create record from domain types
+    let _record = RelayLabelRecord::new(relay_id, &relay_label);
+
+    // Save using repository
+    repo.save_label(relay_id, relay_label.clone())
+        .await
+        .expect("save_label should succeed");
+
+    // Retrieve and verify conversion
+    let retrieved = repo
+        .get_label(relay_id)
+        .await
+        .expect("get_label should succeed");
+
+    assert!(retrieved.is_some(), "Label should be retrieved");
+    assert_eq!(retrieved.unwrap(), relay_label, "Labels should match");
+}
+
+/// Test that repository handles database errors gracefully.
+#[tokio::test]
+async fn test_repository_error_handling() {
+    let _repo = SqliteRelayLabelRepository::in_memory()
+        .await
+        .expect("Failed to create repository");
+
+    // Test with invalid relay ID (should be caught by domain validation)
+    let invalid_relay_id = RelayId::new(9); // This will fail validation
+    assert!(invalid_relay_id.is_err(), "Invalid relay ID should fail validation");
+
+    // Test with invalid label (should be caught by domain validation)
+    let invalid_label = RelayLabel::new("".to_string()); // Empty label
+    assert!(invalid_label.is_err(), "Empty label should fail validation");
+}
+
+/// Test that repository operations are thread-safe.
+#[tokio::test]
+async fn test_concurrent_operations_are_thread_safe() {
+    let repo = SqliteRelayLabelRepository::in_memory()
+        .await
+        .expect("Failed to create repository");
+
+    // Since SqliteRelayLabelRepository doesn't implement Clone, we'll test
+    // sequential operations which still verify the repository handles
+    // multiple operations correctly
+    
+    // Save multiple labels sequentially
+    let relay_id1 = RelayId::new(1).expect("Valid relay ID");
+    let label1 = RelayLabel::new("Task1".to_string()).expect("Valid label");
+    repo.save_label(relay_id1, label1)
+        .await
+        .expect("First save should succeed");
+
+    let relay_id2 = RelayId::new(2).expect("Valid relay ID");
+    let label2 = RelayLabel::new("Task2".to_string()).expect("Valid label");
+    repo.save_label(relay_id2, label2)
+        .await
+        .expect("Second save should succeed");
+
+    let relay_id3 = RelayId::new(3).expect("Valid relay ID");
+    let label3 = RelayLabel::new("Task3".to_string()).expect("Valid label");
+    repo.save_label(relay_id3, label3)
+        .await
+        .expect("Third save should succeed");
+
+    // Verify all labels were saved
+    let all_labels = repo
+        .get_all_labels()
+        .await
+        .expect("get_all_labels should succeed");
+    assert_eq!(all_labels.len(), 3, "Should have all 3 labels");
+}

justfile

@@ -31,7 +31,10 @@ release-run:
 	cargo run --release
 
 test:
-	cargo test
+	cargo test --all --all-targets
+
+test-hardware:
+	cargo test --all --all-targets -- --ignored
 
 coverage:
 	mkdir -p coverage