Fix ZMQ sender blocking - move to independent thread with try_read

CRITICAL FIX: The previous cached collector architecture still had ZMQ sending
in the main event loop, where it could block waiting for RwLock when collectors
were writing. This caused the 3-8 second delays you observed.

Changes:
- Move ZMQ publisher to dedicated std::thread (ZMQ sockets aren't thread-safe)
- Use try_read() instead of read() to avoid blocking on write locks
- Send previous data if cache is locked by collector
- ZMQ now sends every 2s regardless of collector timing
- Remove publisher from ZmqHandler (now only handles commands)

Architecture:
- Collectors: Independent tokio tasks updating shared cache
- ZMQ Sender: Dedicated OS thread with its own publisher socket
- Main Loop: Only handles commands and notifications

This ensures ZMQ transmission is NEVER blocked by slow collectors.

Bump version to v0.1.195

CLAUDE.md

@@ -156,6 +156,86 @@ Complete migration from string-based metrics to structured JSON data. Eliminates
 - ✅ Backward compatibility via bridge conversion to existing UI widgets
 - ✅ All string parsing bugs eliminated
 
+### Cached Collector Architecture (✅ IMPLEMENTED)
+
+**Problem:** Blocking collectors prevent timely ZMQ transmission, causing false "host offline" alerts.
+
+**Previous (Sequential Blocking):**
+```
+Every 1 second:
+  └─ collect_all_data() [BLOCKS for 2-10+ seconds]
+      ├─ CPU (fast: 10ms)
+      ├─ Memory (fast: 20ms)
+      ├─ Disk SMART (slow: 3s per drive × 4 drives = 12s)
+      ├─ Service disk usage (slow: 2-8s per service)
+      └─ Docker (medium: 500ms)
+  └─ send_via_zmq()  [Only after ALL collection completes]
+
+Result: If any collector takes >10s → "host offline" false alert
+```
+
+**New (Cached Independent Collectors):**
+```
+Shared Cache: Arc<RwLock<AgentData>>
+
+Background Collectors (independent async tasks):
+├─ Fast collectors (CPU, RAM, Network)
+│   └─ Update cache every 1 second
+├─ Medium collectors (Services, Docker)
+│   └─ Update cache every 5 seconds
+└─ Slow collectors (Disk usage, SMART data)
+    └─ Update cache every 60 seconds
+
+ZMQ Sender (separate async task):
+Every 1 second:
+  └─ Read current cache
+  └─ Send via ZMQ [Always instant, never blocked]
+```
+
+**Benefits:**
+- ✅ ZMQ sends every 1 second regardless of collector speed
+- ✅ No false "host offline" alerts from slow collectors
+- ✅ Different update rates for different metrics (CPU=1s, SMART=60s)
+- ✅ System stays responsive even with slow operations
+- ✅ Slow collectors can use longer timeouts without blocking
+
+**Implementation Details:**
+- **Shared cache**: `Arc<RwLock<AgentData>>` initialized at agent startup
+- **Collector intervals**: Fully configurable via NixOS config (`interval_seconds` per collector)
+  - Recommended: Fast (1-10s): CPU, Memory, Network
+  - Recommended: Medium (30-60s): Backup, NixOS
+  - Recommended: Slow (60-300s): Disk, Systemd
+- **Independent tasks**: Each collector spawned as separate tokio task in `Agent::new()`
+- **Cache updates**: Collectors acquire write lock → update → release immediately
+- **ZMQ sender**: Main loop reads cache every `collection_interval_seconds` and broadcasts
+- **Notification check**: Runs every `notifications.check_interval_seconds`
+- **Lock strategy**: Short-lived write locks prevent blocking, read locks for transmission
+- **Stale data**: Acceptable for slow-changing metrics (SMART data, disk usage)
+
+**Configuration (NixOS):**
+All intervals and timeouts configurable in `services/cm-dashboard.nix`:
+
+Collection Intervals:
+- `collectors.cpu.interval_seconds` (default: 10s)
+- `collectors.memory.interval_seconds` (default: 2s)
+- `collectors.disk.interval_seconds` (default: 300s)
+- `collectors.systemd.interval_seconds` (default: 10s)
+- `collectors.backup.interval_seconds` (default: 60s)
+- `collectors.network.interval_seconds` (default: 10s)
+- `collectors.nixos.interval_seconds` (default: 60s)
+- `notifications.check_interval_seconds` (default: 30s)
+- `collection_interval_seconds` - ZMQ transmission rate (default: 2s)
+
+Command Timeouts (prevent resource leaks from hung commands):
+- `collectors.disk.command_timeout_seconds` (default: 30s) - lsblk, smartctl, etc.
+- `collectors.systemd.command_timeout_seconds` (default: 15s) - systemctl, docker, du
+- `collectors.network.command_timeout_seconds` (default: 10s) - ip route, ip addr
+
+**Code Locations:**
+- agent/src/agent.rs:59-133 - Collector task spawning
+- agent/src/agent.rs:151-179 - Independent collector task runner
+- agent/src/agent.rs:199-207 - ZMQ sender in main loop
+
 ### Maintenance Mode
 
 - Agent checks for `/tmp/cm-maintenance` file before sending notifications

Cargo.lock

@@ -279,7 +279,7 @@ checksum = "a1d728cc89cf3aee9ff92b05e62b19ee65a02b5702cff7d5a377e32c6ae29d8d"
 
 [[package]]
 name = "cm-dashboard"
-version = "0.1.175"
+version = "0.1.194"
 dependencies = [
  "anyhow",
  "chrono",
@@ -301,7 +301,7 @@ dependencies = [
 
 [[package]]
 name = "cm-dashboard-agent"
-version = "0.1.175"
+version = "0.1.194"
 dependencies = [
  "anyhow",
  "async-trait",
@@ -324,7 +324,7 @@ dependencies = [
 
 [[package]]
 name = "cm-dashboard-shared"
-version = "0.1.175"
+version = "0.1.194"
 dependencies = [
  "chrono",
  "serde",

agent/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "cm-dashboard-agent"
-version = "0.1.178"
+version = "0.1.195"
 edition = "2021"
 
 [dependencies]

agent/src/agent.rs

@@ -1,13 +1,14 @@
 use anyhow::Result;
 use gethostname::gethostname;
+use std::sync::Arc;
 use std::time::Duration;
+use tokio::sync::RwLock;
 use tokio::time::interval;
 use tracing::{debug, error, info};
 
 use crate::communication::{AgentCommand, ZmqHandler};
 use crate::config::AgentConfig;
 use crate::collectors::{
-    Collector,
     backup::BackupCollector,
     cpu::CpuCollector,
     disk::DiskCollector,
@@ -23,7 +24,7 @@ pub struct Agent {
     hostname: String,
     config: AgentConfig,
     zmq_handler: ZmqHandler,
-    collectors: Vec<Box<dyn Collector>>,
+    cache: Arc<RwLock<AgentData>>,
     notification_manager: NotificationManager,
     previous_status: Option<SystemStatus>,
 }
@@ -55,39 +56,94 @@ impl Agent {
             config.zmq.publisher_port
         );
 
-        // Initialize collectors
-        let mut collectors: Vec<Box<dyn Collector>> = Vec::new();
-        
-        // Add enabled collectors
+        // Initialize shared cache
+        let cache = Arc::new(RwLock::new(AgentData::new(
+            hostname.clone(),
+            env!("CARGO_PKG_VERSION").to_string()
+        )));
+        info!("Initialized shared agent data cache");
+
+        // Spawn independent collector tasks
+        let mut collector_count = 0;
+
+        // CPU collector
         if config.collectors.cpu.enabled {
-            collectors.push(Box::new(CpuCollector::new(config.collectors.cpu.clone())));
+            let cache_clone = cache.clone();
+            let collector = CpuCollector::new(config.collectors.cpu.clone());
+            let interval = config.collectors.cpu.interval_seconds;
+            tokio::spawn(async move {
+                Self::run_collector_task(cache_clone, collector, Duration::from_secs(interval), "CPU").await;
+            });
+            collector_count += 1;
         }
-        
+
+        // Memory collector
         if config.collectors.memory.enabled {
-            collectors.push(Box::new(MemoryCollector::new(config.collectors.memory.clone())));
-        }
-        
-        if config.collectors.disk.enabled {
-            collectors.push(Box::new(DiskCollector::new(config.collectors.disk.clone())));
-        }
-        
-        if config.collectors.systemd.enabled {
-            collectors.push(Box::new(SystemdCollector::new(config.collectors.systemd.clone())));
-        }
-        
-        if config.collectors.backup.enabled {
-            collectors.push(Box::new(BackupCollector::new()));
+            let cache_clone = cache.clone();
+            let collector = MemoryCollector::new(config.collectors.memory.clone());
+            let interval = config.collectors.memory.interval_seconds;
+            tokio::spawn(async move {
+                Self::run_collector_task(cache_clone, collector, Duration::from_secs(interval), "Memory").await;
+            });
+            collector_count += 1;
         }
 
+        // Network collector
         if config.collectors.network.enabled {
-            collectors.push(Box::new(NetworkCollector::new(config.collectors.network.clone())));
+            let cache_clone = cache.clone();
+            let collector = NetworkCollector::new(config.collectors.network.clone());
+            let interval = config.collectors.network.interval_seconds;
+            tokio::spawn(async move {
+                Self::run_collector_task(cache_clone, collector, Duration::from_secs(interval), "Network").await;
+            });
+            collector_count += 1;
         }
 
+        // Backup collector
+        if config.collectors.backup.enabled {
+            let cache_clone = cache.clone();
+            let collector = BackupCollector::new();
+            let interval = config.collectors.backup.interval_seconds;
+            tokio::spawn(async move {
+                Self::run_collector_task(cache_clone, collector, Duration::from_secs(interval), "Backup").await;
+            });
+            collector_count += 1;
+        }
+
+        // NixOS collector
         if config.collectors.nixos.enabled {
-            collectors.push(Box::new(NixOSCollector::new(config.collectors.nixos.clone())));
+            let cache_clone = cache.clone();
+            let collector = NixOSCollector::new(config.collectors.nixos.clone());
+            let interval = config.collectors.nixos.interval_seconds;
+            tokio::spawn(async move {
+                Self::run_collector_task(cache_clone, collector, Duration::from_secs(interval), "NixOS").await;
+            });
+            collector_count += 1;
         }
 
-        info!("Initialized {} collectors", collectors.len());
+        // Disk collector
+        if config.collectors.disk.enabled {
+            let cache_clone = cache.clone();
+            let collector = DiskCollector::new(config.collectors.disk.clone());
+            let interval = config.collectors.disk.interval_seconds;
+            tokio::spawn(async move {
+                Self::run_collector_task(cache_clone, collector, Duration::from_secs(interval), "Disk").await;
+            });
+            collector_count += 1;
+        }
+
+        // Systemd collector
+        if config.collectors.systemd.enabled {
+            let cache_clone = cache.clone();
+            let collector = SystemdCollector::new(config.collectors.systemd.clone());
+            let interval = config.collectors.systemd.interval_seconds;
+            tokio::spawn(async move {
+                Self::run_collector_task(cache_clone, collector, Duration::from_secs(interval), "Systemd").await;
+            });
+            collector_count += 1;
+        }
+
+        info!("Spawned {} independent collector tasks", collector_count);
 
         // Initialize notification manager
         let notification_manager = NotificationManager::new(&config.notifications, &hostname)?;
@@ -97,45 +153,121 @@ impl Agent {
             hostname,
             config,
             zmq_handler,
-            collectors,
+            cache,
             notification_manager,
             previous_status: None,
         })
     }
 
-    /// Main agent loop with structured data collection
-    pub async fn run(&mut self, mut shutdown_rx: tokio::sync::oneshot::Receiver<()>) -> Result<()> {
-        info!("Starting agent main loop");
+    /// Independent collector task runner
+    async fn run_collector_task<C>(
+        cache: Arc<RwLock<AgentData>>,
+        collector: C,
+        interval_duration: Duration,
+        name: &str,
+    ) where
+        C: crate::collectors::Collector + Send + 'static,
+    {
+        let mut interval_timer = interval(interval_duration);
+        info!("{} collector task started (interval: {:?})", name, interval_duration);
 
-        // Initial collection
-        if let Err(e) = self.collect_and_broadcast().await {
-            error!("Initial metric collection failed: {}", e);
+        loop {
+            interval_timer.tick().await;
+
+            // Acquire write lock and update cache
+            {
+                let mut agent_data = cache.write().await;
+                match collector.collect_structured(&mut *agent_data).await {
+                    Ok(_) => {
+                        debug!("{} collector updated cache", name);
+                    }
+                    Err(e) => {
+                        error!("{} collector failed: {}", name, e);
+                    }
+                }
+            } // Release lock immediately after collection
         }
+    }
 
-        // Set up intervals
-        let mut transmission_interval = interval(Duration::from_secs(
-            self.config.collection_interval_seconds,
+    /// Main agent loop with cached data architecture
+    pub async fn run(&mut self, mut shutdown_rx: tokio::sync::oneshot::Receiver<()>) -> Result<()> {
+        info!("Starting agent main loop with cached collector architecture");
+
+        // Spawn independent ZMQ sender task
+        // Create dedicated ZMQ publisher for the sender task
+        let cache_clone = self.cache.clone();
+        let publisher_config = self.config.zmq.clone();
+        let transmission_interval_secs = self.config.collection_interval_seconds;
+
+        std::thread::spawn(move || {
+            // Create ZMQ publisher in this thread (ZMQ sockets are not thread-safe)
+            let context = zmq::Context::new();
+            let publisher = context.socket(zmq::SocketType::PUB).unwrap();
+            let bind_address = format!("tcp://{}:{}", publisher_config.bind_address, publisher_config.publisher_port);
+            publisher.bind(&bind_address).unwrap();
+            publisher.set_sndhwm(1000).unwrap();
+            publisher.set_linger(1000).unwrap();
+            info!("ZMQ sender task started on {} (interval: {}s)", bind_address, transmission_interval_secs);
+
+            let mut last_sent_data: Option<AgentData> = None;
+            let interval_duration = std::time::Duration::from_secs(transmission_interval_secs);
+            let mut next_send = std::time::Instant::now() + interval_duration;
+
+            loop {
+                // Sleep until next send time
+                std::thread::sleep(next_send.saturating_duration_since(std::time::Instant::now()));
+                next_send = std::time::Instant::now() + interval_duration;
+
+                // Try to read cache without blocking - if locked, send last known data
+                let data_to_send = match cache_clone.try_read() {
+                    Ok(agent_data) => {
+                        let data_clone = agent_data.clone();
+                        drop(agent_data); // Release lock immediately
+                        last_sent_data = Some(data_clone.clone());
+                        Some(data_clone)
+                    }
+                    Err(_) => {
+                        // Lock is held by collector - use last sent data
+                        debug!("Cache locked by collector, sending previous data");
+                        last_sent_data.clone()
+                    }
+                };
+
+                if let Some(data) = data_to_send {
+                    // Publish via ZMQ
+                    if let Ok(envelope) = cm_dashboard_shared::MessageEnvelope::agent_data(data) {
+                        if let Ok(serialized) = serde_json::to_vec(&envelope) {
+                            if let Err(e) = publisher.send(&serialized, 0) {
+                                error!("Failed to send ZMQ message: {}", e);
+                            } else {
+                                debug!("Successfully broadcast agent data");
+                            }
+                        }
+                    }
+                }
+            }
+        });
+
+        // Set up intervals for notifications and commands
+        let mut notification_interval = interval(Duration::from_secs(
+            self.config.notifications.check_interval_seconds,
         ));
-        let mut notification_interval = interval(Duration::from_secs(30)); // Check notifications every 30s
+        let mut command_interval = interval(Duration::from_millis(100));
 
-        // Skip initial ticks to avoid immediate execution
-        transmission_interval.tick().await;
+        // Skip initial ticks
         notification_interval.tick().await;
+        command_interval.tick().await;
 
         loop {
             tokio::select! {
-                _ = transmission_interval.tick() => {
-                    if let Err(e) = self.collect_and_broadcast().await {
-                        error!("Failed to collect and broadcast metrics: {}", e);
+                _ = notification_interval.tick() => {
+                    // Read cache and check for status changes
+                    let agent_data = self.cache.read().await.clone();
+                    if let Err(e) = self.check_status_changes_and_notify(&agent_data).await {
+                        error!("Failed to check status changes: {}", e);
                     }
                 }
-                _ = notification_interval.tick() => {
-                    // Process any pending notifications
-                    // NOTE: With structured data, we might need to implement status tracking differently
-                    // For now, we skip this until status evaluation is migrated
-                }
-                // Handle incoming commands (check periodically)
-                _ = tokio::time::sleep(Duration::from_millis(100)) => {
+                _ = command_interval.tick() => {
                     if let Err(e) = self.handle_commands().await {
                         error!("Error handling commands: {}", e);
                     }
@@ -151,35 +283,6 @@ impl Agent {
         Ok(())
     }
 
-    /// Collect structured data from all collectors and broadcast via ZMQ
-    async fn collect_and_broadcast(&mut self) -> Result<()> {
-        debug!("Starting structured data collection");
-
-        // Initialize empty AgentData
-        let mut agent_data = AgentData::new(self.hostname.clone(), env!("CARGO_PKG_VERSION").to_string());
-
-        // Collect data from all collectors
-        for collector in &self.collectors {
-            if let Err(e) = collector.collect_structured(&mut agent_data).await {
-                error!("Collector failed: {}", e);
-                // Continue with other collectors even if one fails
-            }
-        }
-
-        // Check for status changes and send notifications
-        if let Err(e) = self.check_status_changes_and_notify(&agent_data).await {
-            error!("Failed to check status changes: {}", e);
-        }
-
-        // Broadcast the structured data via ZMQ
-        if let Err(e) = self.zmq_handler.publish_agent_data(&agent_data).await {
-            error!("Failed to broadcast agent data: {}", e);
-        } else {
-            debug!("Successfully broadcast structured agent data");
-        }
-
-        Ok(())
-    }
 
     /// Check for status changes and send notifications
     async fn check_status_changes_and_notify(&mut self, agent_data: &AgentData) -> Result<()> {
@@ -267,10 +370,10 @@ impl Agent {
 
             match command {
                 AgentCommand::CollectNow => {
-                    info!("Received immediate collection request");
-                    if let Err(e) = self.collect_and_broadcast().await {
-                        error!("Failed to collect on demand: {}", e);
-                    }
+                    info!("Received immediate transmission request");
+                    // With cached architecture and dedicated ZMQ sender thread,
+                    // data is already being sent every interval
+                    // This command is acknowledged but not actionable in new architecture
                 }
                 AgentCommand::SetInterval { seconds } => {
                     info!("Received interval change request: {}s", seconds);

agent/src/collectors/disk.rs

@@ -112,9 +112,12 @@ impl DiskCollector {
 
     /// Get block devices and their mount points using lsblk
     async fn get_mount_devices(&self) -> Result<HashMap<String, String>, CollectorError> {
-        let output = Command::new("lsblk")
-            .args(&["-rn", "-o", "NAME,MOUNTPOINT"])
-            .output()
+        use super::run_command_with_timeout;
+
+        let mut cmd = Command::new("lsblk");
+        cmd.args(&["-rn", "-o", "NAME,MOUNTPOINT"]);
+
+        let output = run_command_with_timeout(cmd, self.config.command_timeout_seconds).await
             .map_err(|e| CollectorError::SystemRead {
                 path: "block devices".to_string(),
                 error: e.to_string(),
@@ -186,8 +189,8 @@ impl DiskCollector {
 
     /// Get filesystem info for a single mount point
     fn get_filesystem_info(&self, mount_point: &str) -> Result<(u64, u64), CollectorError> {
-        let output = Command::new("df")
-            .args(&["--block-size=1", mount_point])
+        let output = std::process::Command::new("timeout")
+            .args(&["2", "df", "--block-size=1", mount_point])
             .output()
             .map_err(|e| CollectorError::SystemRead {
                 path: format!("df {}", mount_point),
@@ -386,7 +389,7 @@ impl DiskCollector {
     /// Get SMART data for drives
     async fn get_smart_data_for_drives(&self, physical_drives: &[PhysicalDrive], mergerfs_pools: &[MergerfsPool]) -> HashMap<String, SmartData> {
         let mut smart_data = HashMap::new();
-        
+
         // Collect all drive names
         let mut all_drives = std::collections::HashSet::new();
         for drive in physical_drives {
@@ -413,23 +416,26 @@ impl DiskCollector {
 
     /// Get SMART data for a single drive
     async fn get_smart_data(&self, drive_name: &str) -> Result<SmartData, CollectorError> {
-        let output = Command::new("sudo")
-            .args(&["smartctl", "-a", &format!("/dev/{}", drive_name)])
-            .output()
+        use super::run_command_with_timeout;
+
+        // Use direct smartctl (no sudo) - service has CAP_SYS_RAWIO and CAP_SYS_ADMIN capabilities
+        // For NVMe drives, specify device type explicitly
+        let mut cmd = Command::new("smartctl");
+        if drive_name.starts_with("nvme") {
+            cmd.args(&["-d", "nvme", "-a", &format!("/dev/{}", drive_name)]);
+        } else {
+            cmd.args(&["-a", &format!("/dev/{}", drive_name)]);
+        }
+
+        let output = run_command_with_timeout(cmd, 3).await
             .map_err(|e| CollectorError::SystemRead {
                 path: format!("SMART data for {}", drive_name),
                 error: e.to_string(),
             })?;
 
         let output_str = String::from_utf8_lossy(&output.stdout);
-        let error_str = String::from_utf8_lossy(&output.stderr);
-        
-        // Debug logging for SMART command results
-        debug!("SMART output for {}: status={}, stdout_len={}, stderr={}", 
-            drive_name, output.status, output_str.len(), error_str);
-        
+
         if !output.status.success() {
-            debug!("SMART command failed for {}: {}", drive_name, error_str);
             // Return unknown data rather than failing completely
             return Ok(SmartData {
                 health: "UNKNOWN".to_string(),
@@ -524,6 +530,9 @@ impl DiskCollector {
 
     /// Populate drives data into AgentData
     fn populate_drives_data(&self, physical_drives: &[PhysicalDrive], smart_data: &HashMap<String, SmartData>, agent_data: &mut AgentData) -> Result<(), CollectorError> {
+        // Clear existing drives data to prevent duplicates in cached architecture
+        agent_data.system.storage.drives.clear();
+
         for drive in physical_drives {
             let smart = smart_data.get(&drive.name);
             
@@ -561,6 +570,9 @@ impl DiskCollector {
 
     /// Populate pools data into AgentData
     fn populate_pools_data(&self, mergerfs_pools: &[MergerfsPool], smart_data: &HashMap<String, SmartData>, agent_data: &mut AgentData) -> Result<(), CollectorError> {
+        // Clear existing pools data to prevent duplicates in cached architecture
+        agent_data.system.storage.pools.clear();
+
         for pool in mergerfs_pools {
             // Calculate pool health and statuses based on member drive health
             let (pool_health, health_status, usage_status, data_drive_data, parity_drive_data) = self.calculate_pool_health(pool, smart_data);
@@ -756,9 +768,9 @@ impl DiskCollector {
 
     /// Get drive information for a mount path
     fn get_drive_info_for_path(&self, path: &str) -> anyhow::Result<PoolDrive> {
-        // Use lsblk to find the backing device
-        let output = Command::new("lsblk")
-            .args(&["-rn", "-o", "NAME,MOUNTPOINT"])
+        // Use lsblk to find the backing device with timeout
+        let output = Command::new("timeout")
+            .args(&["2", "lsblk", "-rn", "-o", "NAME,MOUNTPOINT"])
             .output()
             .map_err(|e| anyhow::anyhow!("Failed to run lsblk: {}", e))?;
             

agent/src/collectors/memory.rs

@@ -97,20 +97,23 @@ impl MemoryCollector {
 
     /// Populate tmpfs data into AgentData
     async fn populate_tmpfs_data(&self, agent_data: &mut AgentData) -> Result<(), CollectorError> {
+        // Clear existing tmpfs data to prevent duplicates in cached architecture
+        agent_data.system.memory.tmpfs.clear();
+
         // Discover all tmpfs mount points
         let tmpfs_mounts = self.discover_tmpfs_mounts()?;
-        
+
         if tmpfs_mounts.is_empty() {
             debug!("No tmpfs mounts found to monitor");
             return Ok(());
         }
 
-        // Get usage data for all tmpfs mounts at once using df
-        let mut df_args = vec!["df", "--output=target,size,used", "--block-size=1"];
+        // Get usage data for all tmpfs mounts at once using df (with 2 second timeout)
+        let mut df_args = vec!["2", "df", "--output=target,size,used", "--block-size=1"];
         df_args.extend(tmpfs_mounts.iter().map(|s| s.as_str()));
 
-        let df_output = std::process::Command::new(df_args[0])
-            .args(&df_args[1..])
+        let df_output = std::process::Command::new("timeout")
+            .args(&df_args[..])
             .output()
             .map_err(|e| CollectorError::SystemRead {
                 path: "tmpfs mounts".to_string(),

agent/src/collectors/mod.rs

@@ -1,6 +1,8 @@
 use async_trait::async_trait;
 use cm_dashboard_shared::{AgentData};
-
+use std::process::{Command, Output};
+use std::time::Duration;
+use tokio::time::timeout;
 
 pub mod backup;
 pub mod cpu;
@@ -13,6 +15,20 @@ pub mod systemd;
 
 pub use error::CollectorError;
 
+/// Run a command with a timeout to prevent blocking
+pub async fn run_command_with_timeout(mut cmd: Command, timeout_secs: u64) -> std::io::Result<Output> {
+    let timeout_duration = Duration::from_secs(timeout_secs);
+
+    match timeout(timeout_duration, tokio::task::spawn_blocking(move || cmd.output())).await {
+        Ok(Ok(result)) => result,
+        Ok(Err(e)) => Err(std::io::Error::new(std::io::ErrorKind::Other, e)),
+        Err(_) => Err(std::io::Error::new(
+            std::io::ErrorKind::TimedOut,
+            format!("Command timed out after {} seconds", timeout_secs)
+        )),
+    }
+}
+
 
 /// Base trait for all collectors with direct structured data output
 #[async_trait]

agent/src/collectors/network.rs

@@ -8,12 +8,12 @@ use crate::config::NetworkConfig;
 
 /// Network interface collector with physical/virtual classification and link status
 pub struct NetworkCollector {
-    _config: NetworkConfig,
+    config: NetworkConfig,
 }
 
 impl NetworkCollector {
     pub fn new(config: NetworkConfig) -> Self {
-        Self { _config: config }
+        Self { config }
     }
 
     /// Check if interface is physical (not virtual)
@@ -50,8 +50,9 @@ impl NetworkCollector {
     }
 
     /// Get the primary physical interface (the one with default route)
-    fn get_primary_physical_interface() -> Option<String> {
-        match Command::new("ip").args(["route", "show", "default"]).output() {
+    fn get_primary_physical_interface(&self) -> Option<String> {
+        let timeout_str = self.config.command_timeout_seconds.to_string();
+        match Command::new("timeout").args([&timeout_str, "ip", "route", "show", "default"]).output() {
             Ok(output) if output.status.success() => {
                 let output_str = String::from_utf8_lossy(&output.stdout);
                 // Parse: "default via 192.168.1.1 dev eno1 ..."
@@ -110,7 +111,8 @@ impl NetworkCollector {
         // Parse VLAN configuration
         let vlan_map = Self::parse_vlan_config();
 
-        match Command::new("ip").args(["-j", "addr"]).output() {
+        let timeout_str = self.config.command_timeout_seconds.to_string();
+        match Command::new("timeout").args([&timeout_str, "ip", "-j", "addr"]).output() {
             Ok(output) if output.status.success() => {
                 let json_str = String::from_utf8_lossy(&output.stdout);
 
@@ -195,7 +197,7 @@ impl NetworkCollector {
         }
 
         // Assign primary physical interface as parent to virtual interfaces without explicit parent
-        let primary_interface = Self::get_primary_physical_interface();
+        let primary_interface = self.get_primary_physical_interface();
         if let Some(primary) = primary_interface {
             for interface in interfaces.iter_mut() {
                 // Only assign parent to virtual interfaces that don't already have one

agent/src/collectors/nixos.rs

@@ -43,8 +43,8 @@ impl NixOSCollector {
         match fs::read_to_string("/etc/hostname") {
             Ok(hostname) => Some(hostname.trim().to_string()),
             Err(_) => {
-                // Fallback to hostname command
-                match Command::new("hostname").output() {
+                // Fallback to hostname command (with 2 second timeout)
+                match Command::new("timeout").args(["2", "hostname"]).output() {
                     Ok(output) => Some(String::from_utf8_lossy(&output.stdout).trim().to_string()),
                     Err(_) => None,
                 }

agent/src/collectors/systemd.rs

@@ -4,7 +4,7 @@ use cm_dashboard_shared::{AgentData, ServiceData, SubServiceData, SubServiceMetr
 use std::process::Command;
 use std::sync::RwLock;
 use std::time::Instant;
-use tracing::{debug, info};
+use tracing::{debug, warn};
 
 use super::{Collector, CollectorError};
 use crate::config::SystemdConfig;
@@ -113,15 +113,13 @@ impl SystemdCollector {
                                 name: site_name.clone(),
                                 service_status: self.calculate_service_status(&site_name, &site_status),
                                 metrics,
+                                service_type: "nginx_site".to_string(),
                             });
                         }
                     }
 
                     if service_name.contains("docker") && active_status == "active" {
-                        info!("Collecting Docker sub-services for service: {}", service_name);
-
                         let docker_containers = self.get_docker_containers();
-                        info!("Found {} Docker containers", docker_containers.len());
                         for (container_name, container_status) in docker_containers {
                             // For now, docker containers have no additional metrics
                             // Future: could add memory_mb, cpu_percent, restart_count, etc.
@@ -131,28 +129,27 @@ impl SystemdCollector {
                                 name: container_name.clone(),
                                 service_status: self.calculate_service_status(&container_name, &container_status),
                                 metrics,
+                                service_type: "container".to_string(),
                             });
                         }
 
                         // Add Docker images
                         let docker_images = self.get_docker_images();
-                        info!("Found {} Docker images", docker_images.len());
-                        for (image_name, image_status, image_size) in docker_images {
+                        for (image_name, image_status, image_size_mb) in docker_images {
                             let mut metrics = Vec::new();
                             metrics.push(SubServiceMetric {
                                 label: "size".to_string(),
-                                value: 0.0, // Size as string in name instead
-                                unit: None,
+                                value: image_size_mb,
+                                unit: Some("MB".to_string()),
                             });
 
                             sub_services.push(SubServiceData {
-                                name: format!("{} ({})", image_name, image_size),
+                                name: image_name.to_string(),
                                 service_status: self.calculate_service_status(&image_name, &image_status),
                                 metrics,
+                                service_type: "image".to_string(),
                             });
                         }
-
-                        info!("Total Docker sub-services added: {}", sub_services.len());
                     }
 
                     // Create complete service data
@@ -175,6 +172,10 @@ impl SystemdCollector {
             }
         }
 
+        // Sort services alphabetically by name
+        agent_data.services.sort_by(|a, b| a.name.cmp(&b.name));
+        complete_service_data.sort_by(|a, b| a.name.cmp(&b.name));
+
         // Update cached state
         {
             let mut state = self.state.write().unwrap();
@@ -254,8 +255,9 @@ impl SystemdCollector {
     /// Auto-discover interesting services to monitor
     fn discover_services_internal(&self) -> Result<(Vec<String>, std::collections::HashMap<String, ServiceStatusInfo>)> {
         // First: Get all service unit files
-        let unit_files_output = Command::new("systemctl")
-            .args(&["list-unit-files", "--type=service", "--no-pager", "--plain"])
+        let timeout_str = self.config.command_timeout_seconds.to_string();
+        let unit_files_output = Command::new("timeout")
+            .args(&[&timeout_str, "systemctl", "list-unit-files", "--type=service", "--no-pager", "--plain"])
             .output()?;
 
         if !unit_files_output.status.success() {
@@ -263,8 +265,8 @@ impl SystemdCollector {
         }
 
         // Second: Get runtime status of all units
-        let units_status_output = Command::new("systemctl")
-            .args(&["list-units", "--type=service", "--all", "--no-pager", "--plain"])
+        let units_status_output = Command::new("timeout")
+            .args(&[&timeout_str, "systemctl", "list-units", "--type=service", "--all", "--no-pager", "--plain"])
             .output()?;
 
         if !units_status_output.status.success() {
@@ -361,15 +363,16 @@ impl SystemdCollector {
         }
 
         // Fallback to systemctl if not in cache
-        let output = Command::new("systemctl")
-            .args(&["is-active", &format!("{}.service", service)])
+        let timeout_str = self.config.command_timeout_seconds.to_string();
+        let output = Command::new("timeout")
+            .args(&[&timeout_str, "systemctl", "is-active", &format!("{}.service", service)])
             .output()?;
 
         let active_status = String::from_utf8(output.stdout)?.trim().to_string();
 
         // Get more detailed info
-        let output = Command::new("systemctl")
-            .args(&["show", &format!("{}.service", service), "--property=LoadState,ActiveState,SubState"])
+        let output = Command::new("timeout")
+            .args(&[&timeout_str, "systemctl", "show", &format!("{}.service", service), "--property=LoadState,ActiveState,SubState"])
             .output()?;
 
         let detailed_info = String::from_utf8(output.stdout)?;
@@ -421,7 +424,7 @@ impl SystemdCollector {
         if let Some(dirs) = self.config.service_directories.get(service_name) {
             // Service has configured paths - use the first accessible one
             for dir in dirs {
-                if let Some(size) = self.get_directory_size(dir) {
+                if let Some(size) = self.get_directory_size(dir).await {
                     return Ok(size);
                 }
             }
@@ -430,8 +433,9 @@ impl SystemdCollector {
         }
 
         // No configured path - try to get WorkingDirectory from systemctl
-        let output = Command::new("systemctl")
-            .args(&["show", &format!("{}.service", service_name), "--property=WorkingDirectory"])
+        let timeout_str = self.config.command_timeout_seconds.to_string();
+        let output = Command::new("timeout")
+            .args(&[&timeout_str, "systemctl", "show", &format!("{}.service", service_name), "--property=WorkingDirectory"])
             .output()
             .map_err(|e| CollectorError::SystemRead {
                 path: format!("WorkingDirectory for {}", service_name),
@@ -443,7 +447,7 @@ impl SystemdCollector {
             if line.starts_with("WorkingDirectory=") && !line.contains("[not set]") {
                 let dir = line.strip_prefix("WorkingDirectory=").unwrap_or("");
                 if !dir.is_empty() && dir != "/" {
-                    return Ok(self.get_directory_size(dir).unwrap_or(0.0));
+                    return Ok(self.get_directory_size(dir).await.unwrap_or(0.0));
                 }
             }
         }
@@ -452,17 +456,22 @@ impl SystemdCollector {
     }
     
     /// Get size of a directory in GB
-    fn get_directory_size(&self, path: &str) -> Option<f32> {
-        let output = Command::new("sudo")
-            .args(&["du", "-sb", path])
-            .output()
-            .ok()?;
+    async fn get_directory_size(&self, path: &str) -> Option<f32> {
+        use super::run_command_with_timeout;
+
+        // Use -s (summary) and --apparent-size for speed
+        let mut cmd = Command::new("sudo");
+        cmd.args(&["du", "-s", "--apparent-size", "--block-size=1", path]);
+
+        let output = run_command_with_timeout(cmd, self.config.command_timeout_seconds).await.ok()?;
 
         if !output.status.success() {
             // Log permission errors for debugging but don't spam logs
             let stderr = String::from_utf8_lossy(&output.stderr);
             if stderr.contains("Permission denied") {
                 debug!("Permission denied accessing directory: {}", path);
+            } else if stderr.contains("timed out") {
+                warn!("Directory size check timed out for {}", path);
             } else {
                 debug!("Failed to get size for directory {}: {}", path, stderr);
             }
@@ -781,8 +790,9 @@ impl SystemdCollector {
 
         // Check if docker is available (cm-agent user is in docker group)
         // Use -a to show ALL containers (running and stopped)
-        let output = Command::new("docker")
-            .args(&["ps", "-a", "--format", "{{.Names}},{{.Status}}"])
+        let timeout_str = self.config.command_timeout_seconds.to_string();
+        let output = Command::new("timeout")
+            .args(&[&timeout_str, "docker", "ps", "-a", "--format", "{{.Names}},{{.Status}}"])
             .output();
 
         let output = match output {
@@ -821,25 +831,20 @@ impl SystemdCollector {
     }
 
     /// Get docker images as sub-services
-    fn get_docker_images(&self) -> Vec<(String, String, String)> {
+    fn get_docker_images(&self) -> Vec<(String, String, f32)> {
         let mut images = Vec::new();
-
-        info!("Collecting Docker images");
-
         // Check if docker is available (cm-agent user is in docker group)
-        let output = Command::new("docker")
-            .args(&["images", "--format", "{{.Repository}}:{{.Tag}},{{.Size}}"])
+        let timeout_str = self.config.command_timeout_seconds.to_string();
+        let output = Command::new("timeout")
+            .args(&[&timeout_str, "docker", "images", "--format", "{{.Repository}}:{{.Tag}},{{.Size}}"])
             .output();
 
         let output = match output {
             Ok(out) if out.status.success() => out,
-            Ok(out) => {
-                let stderr = String::from_utf8_lossy(&out.stderr);
-                info!("Docker images command failed with status: {}, stderr: {}", out.status, stderr);
+            Ok(_) => {
                 return images;
             }
-            Err(e) => {
-                info!("Docker images command error: {}", e);
+            Err(_) => {
                 return images;
             }
         };
@@ -857,28 +862,62 @@ impl SystemdCollector {
             let parts: Vec<&str> = line.split(',').collect();
             if parts.len() >= 2 {
                 let image_name = parts[0].trim();
-                let size = parts[1].trim();
+                let size_str = parts[1].trim();
 
                 // Skip <none>:<none> images (dangling images)
                 if image_name.contains("<none>") {
                     continue;
                 }
 
+                // Parse size to MB (sizes come as "142MB", "1.5GB", "512kB", etc.)
+                let size_mb = self.parse_docker_size(size_str);
+
                 images.push((
-                    format!("image_{}", image_name),
-                    "active".to_string(), // Images are always "active" (present)
-                    size.to_string()
+                    image_name.to_string(),
+                    "inactive".to_string(), // Images are informational - use inactive for neutral display
+                    size_mb
                 ));
             }
         }
 
         images
     }
+
+    /// Parse Docker size string to MB
+    fn parse_docker_size(&self, size_str: &str) -> f32 {
+        let size_upper = size_str.to_uppercase();
+
+        // Extract numeric part and unit
+        let mut num_str = String::new();
+        let mut unit = String::new();
+
+        for ch in size_upper.chars() {
+            if ch.is_ascii_digit() || ch == '.' {
+                num_str.push(ch);
+            } else if ch.is_alphabetic() {
+                unit.push(ch);
+            }
+        }
+
+        let value: f32 = num_str.parse().unwrap_or(0.0);
+
+        // Convert to MB
+        match unit.as_str() {
+            "KB" | "K" => value / 1024.0,
+            "MB" | "M" => value,
+            "GB" | "G" => value * 1024.0,
+            "TB" | "T" => value * 1024.0 * 1024.0,
+            _ => value, // Assume bytes if no unit
+        }
+    }
 }
 
 #[async_trait]
 impl Collector for SystemdCollector {
     async fn collect_structured(&self, agent_data: &mut AgentData) -> Result<(), CollectorError> {
+        // Clear existing services data to prevent duplicates in cached architecture
+        agent_data.services.clear();
+
         // Use cached complete data if available and fresh
         if let Some(cached_complete_services) = self.get_cached_complete_services() {
             for service_data in cached_complete_services {

agent/src/communication/mod.rs

@@ -1,13 +1,12 @@
 use anyhow::Result;
-use cm_dashboard_shared::{AgentData, MessageEnvelope};
 use tracing::{debug, info};
 use zmq::{Context, Socket, SocketType};
 
 use crate::config::ZmqConfig;
 
-/// ZMQ communication handler for publishing metrics and receiving commands
+/// ZMQ communication handler for receiving commands
+/// NOTE: Publishing is handled by dedicated thread in Agent::run()
 pub struct ZmqHandler {
-    publisher: Socket,
     command_receiver: Socket,
 }
 
@@ -15,17 +14,6 @@ impl ZmqHandler {
     pub async fn new(config: &ZmqConfig) -> Result<Self> {
         let context = Context::new();
 
-        // Create publisher socket for metrics
-        let publisher = context.socket(SocketType::PUB)?;
-        let pub_bind_address = format!("tcp://{}:{}", config.bind_address, config.publisher_port);
-        publisher.bind(&pub_bind_address)?;
-
-        info!("ZMQ publisher bound to {}", pub_bind_address);
-
-        // Set socket options for efficiency
-        publisher.set_sndhwm(1000)?; // High water mark for outbound messages
-        publisher.set_linger(1000)?; // Linger time on close
-
         // Create command receiver socket (PULL socket to receive commands from dashboard)
         let command_receiver = context.socket(SocketType::PULL)?;
         let cmd_bind_address = format!("tcp://{}:{}", config.bind_address, config.command_port);
@@ -38,33 +26,10 @@ impl ZmqHandler {
         command_receiver.set_linger(1000)?;
 
         Ok(Self {
-            publisher,
             command_receiver,
         })
     }
 
-
-    /// Publish agent data via ZMQ
-    pub async fn publish_agent_data(&self, data: &AgentData) -> Result<()> {
-        debug!(
-            "Publishing agent data for host {}",
-            data.hostname
-        );
-
-        // Create message envelope for agent data
-        let envelope = MessageEnvelope::agent_data(data.clone())
-            .map_err(|e| anyhow::anyhow!("Failed to create agent data envelope: {}", e))?;
-
-        // Serialize envelope
-        let serialized = serde_json::to_vec(&envelope)?;
-
-        // Send via ZMQ
-        self.publisher.send(&serialized, 0)?;
-
-        debug!("Published agent data message ({} bytes)", serialized.len());
-        Ok(())
-    }
-
     /// Try to receive a command (non-blocking)
     pub fn try_receive_command(&self) -> Result<Option<AgentCommand>> {
         match self.command_receiver.recv_bytes(zmq::DONTWAIT) {

agent/src/config/mod.rs

@@ -79,6 +79,9 @@ pub struct DiskConfig {
     pub temperature_critical_celsius: f32,
     pub wear_warning_percent: f32,
     pub wear_critical_percent: f32,
+    /// Command timeout in seconds for lsblk, smartctl, etc.
+    #[serde(default = "default_disk_command_timeout")]
+    pub command_timeout_seconds: u64,
 }
 
 /// Filesystem configuration entry
@@ -108,6 +111,9 @@ pub struct SystemdConfig {
     pub http_timeout_seconds: u64,
     pub http_connect_timeout_seconds: u64,
     pub nginx_latency_critical_ms: f32,
+    /// Command timeout in seconds for systemctl, docker, du commands
+    #[serde(default = "default_systemd_command_timeout")]
+    pub command_timeout_seconds: u64,
 }
 
 
@@ -132,6 +138,9 @@ pub struct BackupConfig {
 pub struct NetworkConfig {
     pub enabled: bool,
     pub interval_seconds: u64,
+    /// Command timeout in seconds for ip route, ip addr commands
+    #[serde(default = "default_network_command_timeout")]
+    pub command_timeout_seconds: u64,
 }
 
 /// Notification configuration
@@ -145,6 +154,9 @@ pub struct NotificationConfig {
     pub rate_limit_minutes: u64,
     /// Email notification batching interval in seconds (default: 60)
     pub aggregation_interval_seconds: u64,
+    /// Status check interval in seconds for detecting changes (default: 30)
+    #[serde(default = "default_notification_check_interval")]
+    pub check_interval_seconds: u64,
     /// List of metric names to exclude from email notifications
     #[serde(default)]
     pub exclude_email_metrics: Vec<String>,
@@ -158,10 +170,26 @@ fn default_heartbeat_interval_seconds() -> u64 {
     5
 }
 
+fn default_notification_check_interval() -> u64 {
+    30
+}
+
 fn default_maintenance_mode_file() -> String {
     "/tmp/cm-maintenance".to_string()
 }
 
+fn default_disk_command_timeout() -> u64 {
+    30
+}
+
+fn default_systemd_command_timeout() -> u64 {
+    15
+}
+
+fn default_network_command_timeout() -> u64 {
+    10
+}
+
 impl AgentConfig {
     pub fn from_file<P: AsRef<Path>>(path: P) -> Result<Self> {
         loader::load_config(path)

dashboard/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "cm-dashboard"
-version = "0.1.178"
+version = "0.1.195"
 edition = "2021"
 
 [dependencies]

dashboard/src/app.rs

@@ -215,7 +215,7 @@ impl Dashboard {
 
                     // Update TUI with new metrics (only if not headless)
                     if let Some(ref mut tui_app) = self.tui_app {
-                        tui_app.update_metrics(&self.metric_store);
+                        tui_app.update_metrics(&mut self.metric_store);
                     }
                 }
                 

dashboard/src/metrics/store.rs

@@ -5,6 +5,14 @@ use tracing::{debug, info, warn};
 
 use super::MetricDataPoint;
 
+/// ZMQ communication statistics per host
+#[derive(Debug, Clone)]
+pub struct ZmqStats {
+    pub packets_received: u64,
+    pub last_packet_time: Instant,
+    pub last_packet_age_secs: f64,
+}
+
 /// Central metric storage for the dashboard
 pub struct MetricStore {
     /// Current structured data: hostname -> AgentData
@@ -13,6 +21,8 @@ pub struct MetricStore {
     historical_metrics: HashMap<String, Vec<MetricDataPoint>>,
     /// Last heartbeat timestamp per host
     last_heartbeat: HashMap<String, Instant>,
+    /// ZMQ communication statistics per host
+    zmq_stats: HashMap<String, ZmqStats>,
     /// Configuration
     max_metrics_per_host: usize,
     history_retention: Duration,
@@ -24,6 +34,7 @@ impl MetricStore {
             current_agent_data: HashMap::new(),
             historical_metrics: HashMap::new(),
             last_heartbeat: HashMap::new(),
+            zmq_stats: HashMap::new(),
             max_metrics_per_host,
             history_retention: Duration::from_secs(history_retention_hours * 3600),
         }
@@ -44,6 +55,16 @@ impl MetricStore {
         self.last_heartbeat.insert(hostname.clone(), now);
         debug!("Updated heartbeat for host {}", hostname);
 
+        // Update ZMQ stats
+        let stats = self.zmq_stats.entry(hostname.clone()).or_insert(ZmqStats {
+            packets_received: 0,
+            last_packet_time: now,
+            last_packet_age_secs: 0.0,
+        });
+        stats.packets_received += 1;
+        stats.last_packet_time = now;
+        stats.last_packet_age_secs = 0.0; // Just received
+
         // Add to history
         let host_history = self
             .historical_metrics
@@ -65,6 +86,15 @@ impl MetricStore {
         self.current_agent_data.get(hostname)
     }
 
+    /// Get ZMQ communication statistics for a host
+    pub fn get_zmq_stats(&mut self, hostname: &str) -> Option<ZmqStats> {
+        let now = Instant::now();
+        self.zmq_stats.get_mut(hostname).map(|stats| {
+            // Update packet age
+            stats.last_packet_age_secs = now.duration_since(stats.last_packet_time).as_secs_f64();
+            stats.clone()
+        })
+    }
 
     /// Get connected hosts (hosts with recent heartbeats)
     pub fn get_connected_hosts(&self, timeout: Duration) -> Vec<String> {

dashboard/src/ui/mod.rs

@@ -100,7 +100,7 @@ impl TuiApp {
     }
 
     /// Update widgets with structured data from store (only for current host)
-    pub fn update_metrics(&mut self, metric_store: &MetricStore) {
+    pub fn update_metrics(&mut self, metric_store: &mut MetricStore) {
         if let Some(hostname) = self.current_host.clone() {
             // Get structured data for this host
             if let Some(agent_data) = metric_store.get_agent_data(&hostname) {
@@ -110,6 +110,14 @@ impl TuiApp {
                 host_widgets.system_widget.update_from_agent_data(agent_data);
                 host_widgets.services_widget.update_from_agent_data(agent_data);
 
+                // Update ZMQ stats
+                if let Some(zmq_stats) = metric_store.get_zmq_stats(&hostname) {
+                    host_widgets.system_widget.update_zmq_stats(
+                        zmq_stats.packets_received,
+                        zmq_stats.last_packet_age_secs
+                    );
+                }
+
                 host_widgets.last_update = Some(Instant::now());
             }
         }

dashboard/src/ui/widgets/services.rs

@@ -32,6 +32,7 @@ struct ServiceInfo {
     disk_gb: Option<f32>,
     metrics: Vec<(String, f32, Option<String>)>, // (label, value, unit)
     widget_status: Status,
+    service_type: String, // "nginx_site", "container", "image", or empty for parent services
 }
 
 impl ServicesWidget {
@@ -169,7 +170,7 @@ impl ServicesWidget {
             // Convert Status enum to display text for sub-services
             match info.widget_status {
                 Status::Ok => "active",
-                Status::Inactive => "inactive", 
+                Status::Inactive => "inactive",
                 Status::Critical => "failed",
                 Status::Pending => "pending",
                 Status::Warning => "warning",
@@ -179,32 +180,62 @@ impl ServicesWidget {
         };
         let tree_symbol = if is_last { "└─" } else { "├─" };
 
-        vec![
-            // Indentation and tree prefix
-            ratatui::text::Span::styled(
-                format!("  {} ", tree_symbol),
-                Typography::tree(),
-            ),
-            // Status icon
-            ratatui::text::Span::styled(
-                format!("{} ", icon),
-                Style::default().fg(status_color).bg(Theme::background()),
-            ),
-            // Service name
-            ratatui::text::Span::styled(
-                format!("{:<18} ", short_name),
-                Style::default()
-                    .fg(Theme::secondary_text())
-                    .bg(Theme::background()),
-            ),
-            // Status/latency text
-            ratatui::text::Span::styled(
-                status_str,
-                Style::default()
-                    .fg(Theme::secondary_text())
-                    .bg(Theme::background()),
-            ),
-        ]
+        // Docker images use docker whale icon
+        if info.service_type == "image" {
+            vec![
+                // Indentation and tree prefix
+                ratatui::text::Span::styled(
+                    format!("  {} ", tree_symbol),
+                    Typography::tree(),
+                ),
+                // Docker icon (simple character for performance)
+                ratatui::text::Span::styled(
+                    "D ".to_string(),
+                    Style::default().fg(Theme::highlight()).bg(Theme::background()),
+                ),
+                // Service name
+                ratatui::text::Span::styled(
+                    format!("{:<18} ", short_name),
+                    Style::default()
+                        .fg(Theme::secondary_text())
+                        .bg(Theme::background()),
+                ),
+                // Status/metrics text
+                ratatui::text::Span::styled(
+                    status_str,
+                    Style::default()
+                        .fg(Theme::secondary_text())
+                        .bg(Theme::background()),
+                ),
+            ]
+        } else {
+            vec![
+                // Indentation and tree prefix
+                ratatui::text::Span::styled(
+                    format!("  {} ", tree_symbol),
+                    Typography::tree(),
+                ),
+                // Status icon
+                ratatui::text::Span::styled(
+                    format!("{} ", icon),
+                    Style::default().fg(status_color).bg(Theme::background()),
+                ),
+                // Service name
+                ratatui::text::Span::styled(
+                    format!("{:<18} ", short_name),
+                    Style::default()
+                        .fg(Theme::secondary_text())
+                        .bg(Theme::background()),
+                ),
+                // Status/latency text
+                ratatui::text::Span::styled(
+                    status_str,
+                    Style::default()
+                        .fg(Theme::secondary_text())
+                        .bg(Theme::background()),
+                ),
+            ]
+        }
     }
 
     /// Move selection up
@@ -282,9 +313,10 @@ impl Widget for ServicesWidget {
                 disk_gb: Some(service.disk_gb),
                 metrics: Vec::new(), // Parent services don't have custom metrics
                 widget_status: service.service_status,
+                service_type: String::new(), // Parent services have no type
             };
             self.parent_services.insert(service.name.clone(), parent_info);
-            
+
             // Process sub-services if any
             if !service.sub_services.is_empty() {
                 let mut sub_list = Vec::new();
@@ -293,12 +325,13 @@ impl Widget for ServicesWidget {
                     let metrics: Vec<(String, f32, Option<String>)> = sub_service.metrics.iter()
                         .map(|m| (m.label.clone(), m.value, m.unit.clone()))
                         .collect();
-                    
+
                     let sub_info = ServiceInfo {
                         memory_mb: None, // Not used for sub-services
                         disk_gb: None,   // Not used for sub-services
                         metrics,
                         widget_status: sub_service.service_status,
+                        service_type: sub_service.service_type.clone(),
                     };
                     sub_list.push((sub_service.name.clone(), sub_info));
                 }
@@ -342,6 +375,7 @@ impl ServicesWidget {
                                     disk_gb: None,
                                     metrics: Vec::new(),
                                     widget_status: Status::Unknown,
+                                    service_type: String::new(),
                                 });
 
                         if metric.name.ends_with("_status") {
@@ -377,6 +411,7 @@ impl ServicesWidget {
                                     disk_gb: None,
                                     metrics: Vec::new(),
                                     widget_status: Status::Unknown,
+                                    service_type: String::new(), // Unknown type in legacy path
                                 },
                             ));
                             &mut sub_service_list.last_mut().unwrap().1

dashboard/src/ui/widgets/system.rs

@@ -15,6 +15,10 @@ pub struct SystemWidget {
     nixos_build: Option<String>,
     agent_hash: Option<String>,
 
+    // ZMQ communication stats
+    zmq_packets_received: Option<u64>,
+    zmq_last_packet_age: Option<f64>,
+
     // Network interfaces
     network_interfaces: Vec<cm_dashboard_shared::NetworkInterfaceData>,
 
@@ -92,6 +96,8 @@ impl SystemWidget {
         Self {
             nixos_build: None,
             agent_hash: None,
+            zmq_packets_received: None,
+            zmq_last_packet_age: None,
             network_interfaces: Vec::new(),
             cpu_load_1min: None,
             cpu_load_5min: None,
@@ -154,6 +160,12 @@ impl SystemWidget {
     pub fn _get_agent_hash(&self) -> Option<&String> {
         self.agent_hash.as_ref()
     }
+
+    /// Update ZMQ communication statistics
+    pub fn update_zmq_stats(&mut self, packets_received: u64, last_packet_age_secs: f64) {
+        self.zmq_packets_received = Some(packets_received);
+        self.zmq_last_packet_age = Some(last_packet_age_secs);
+    }
 }
 
 use super::Widget;
@@ -796,6 +808,18 @@ impl SystemWidget {
             Span::styled(format!("Agent: {}", agent_version_text), Typography::secondary())
         ]));
 
+        // ZMQ communication stats
+        if let (Some(packets), Some(age)) = (self.zmq_packets_received, self.zmq_last_packet_age) {
+            let age_text = if age < 1.0 {
+                format!("{:.0}ms ago", age * 1000.0)
+            } else {
+                format!("{:.1}s ago", age)
+            };
+            lines.push(Line::from(vec![
+                Span::styled(format!("ZMQ: {} pkts, last {}", packets, age_text), Typography::secondary())
+            ]));
+        }
+
         // CPU section
         lines.push(Line::from(vec![
             Span::styled("CPU:", Typography::widget_title())

shared/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "cm-dashboard-shared"
-version = "0.1.178"
+version = "0.1.195"
 edition = "2021"
 
 [dependencies]

shared/src/agent_data.rs

@@ -149,6 +149,9 @@ pub struct SubServiceData {
     pub name: String,
     pub service_status: Status,
     pub metrics: Vec<SubServiceMetric>,
+    /// Type of sub-service: "nginx_site", "container", "image"
+    #[serde(default)]
+    pub service_type: String,
 }
 
 /// Individual metric for a sub-service