Refactor ws_stats.rs to reduce cognitive_complexity and add test coverage

[mukeshblackhat]

ws_stats.rs has two suppressions: run_collector (line 235) and handle_stats_socket (line 364), both #[allow(clippy::cognitive_complexity)].

What they do

• run_collector (~65 lines) — streams container stats via docker.stats() in a loop, extracts CPU/mem/net/block values, pushes to history ring, broadcasts to subscribers, and handles cleanup only if this collector's broadcast sender is still the active one (guards against replacement collectors)
• handle_stats_socket (~55 lines) — replays history buffer to a newly connected WebSocket client, then runs a tokio::select! loop relaying broadcast receiver output to the socket (nearly identical to ws_service_stats.rs and ws_host_service_stats.rs)

What to do

1. Extract stats chunk processing from run_collector

Lines 256–276 process a single stats chunk: extract values, lock history, push to ring, send via broadcast. Extract:

async fn process_stats_chunk(
    state: &AppState,
    key: &str,
    stats: Stats,
    prev_cpu_total: &mut u64,
    prev_cpu_system: &mut u64,
    tx: &broadcast::Sender<serde_json::Value>,
)

2. Extract collector cleanup guard

Lines 281–300 check whether this collector's broadcast sender is still the active one and only remove it if so. Extract:

async fn cleanup_collector_if_current(
    state: &AppState,
    key: &str,
    tx: &broadcast::Sender<serde_json::Value>,
)

3. Extract history replay for handle_stats_socket

Lines 369–379 replay the history ring to the socket. Extract (shared with ws_service_stats.rs and ws_host_service_stats.rs):

async fn replay_stats_history(
    socket: &mut WebSocket,
    history: &Mutex<HashMap<String, VecDeque<serde_json::Value>>>,
    key: &str,
) -> ControlFlow<()>

4. Extract broadcast forwarding loop

The tokio::select! loop that forwards broadcast messages to the socket. Shared across all three stats handlers:

async fn relay_broadcast_to_socket(
    socket: &mut WebSocket,
    rx: &mut broadcast::Receiver<serde_json::Value>,
    key: &str,
)

5. Add unit tests

cleanup_collector_if_current:

• When tx matches current → removes entry
• When tx differs (replacement collector) → leaves entry

Important

The workspace Cargo.toml sets cognitive_complexity, too_many_lines, too_many_arguments, and several other clippy lints to warn globally. Run make lint after every change and make sure the output is clean. Don't fix a suppression here only to create new warnings elsewhere.

Acceptance criteria

• process_stats_chunk extracted (from run_collector)
• cleanup_collector_if_current extracted with unit tests
• replay_stats_history extracted (shared with other stats handlers)
• relay_broadcast_to_socket extracted (shared)
• Both functions use the new helpers
• All tests pass (make test)
• make lint passes with zero new warnings
• Remove both #[allow(clippy::cognitive_complexity)] suppressions if make lint passes without them, keep any still needed (note why)

Related

Tracking issue pattern from previous PRs: #122, #133, #157, #158, #171, #172, #173, #174, #175