fix: resolve PTY failure, MMDS file writes, and metrics instability in envd-rs

Three bugs fixed:

1. PTY connections failed because home directory was hardcoded as
   /home/{username} instead of reading from /etc/passwd. For root,
   this produced /home/root/ which doesn't exist — CWD validation
   rejected every PTY Start request without explicit cwd. Fixed all
   6 locations to use user.dir from nix::unistd::User.

2. MMDS polling silently failed to parse metadata because the
   logs_collector_address field lacked #[serde(default)]. The host
   agent only sends instanceID + envID — missing "address" field
   caused every deserialize attempt to fail, so .WRENN_SANDBOX_ID
   and .WRENN_TEMPLATE_ID were never written. Also added error
   logging and create_dir_all before file writes.

3. Metrics CPU values were non-deterministic because a fresh
   sysinfo::System was created per request with a 100ms sleep
   between reads. Replaced with a background thread that samples
   CPU at fixed 1-second intervals via a persistent System instance,
   matching gopsutil's internal caching behavior. Metrics endpoint
   now reads cached atomic values — no blocking, consistent window.

Also: close master PTY fd in child pre_exec, add process.Start
request logging, bump version to 0.2.0.

envd-rs/src/http/files.rs

@@ -95,7 +95,7 @@ pub async fn get_files(
         Err(e) => return json_error(StatusCode::UNAUTHORIZED, &e),
     };
 
-    let home_dir = format!("/home/{}", user.name);
+    let home_dir = user.dir.to_string_lossy().to_string();
     let resolved = match expand_and_resolve(path_str, &home_dir, state.defaults.workdir.as_deref())
     {
         Ok(p) => p,
@@ -246,7 +246,7 @@ pub async fn post_files(
         Err(e) => return json_error(StatusCode::UNAUTHORIZED, &e),
     };
 
-    let home_dir = format!("/home/{}", user.name);
+    let home_dir = user.dir.to_string_lossy().to_string();
     let uid = user.uid;
     let gid = user.gid;
 

envd-rs/src/http/metrics.rs

@@ -22,10 +22,10 @@ pub struct Metrics {
     disk_total: u64,
 }
 
-pub async fn get_metrics(State(_state): State<Arc<AppState>>) -> impl IntoResponse {
+pub async fn get_metrics(State(state): State<Arc<AppState>>) -> impl IntoResponse {
     tracing::trace!("get metrics");
 
-    match collect_metrics() {
+    match collect_metrics(&state) {
         Ok(m) => (
             StatusCode::OK,
             [(header::CACHE_CONTROL, "no-store")],
@@ -39,26 +39,12 @@ pub async fn get_metrics(State(_state): State<Arc<AppState>>) -> impl IntoRespon
     }
 }
 
-fn collect_metrics() -> Result<Metrics, String> {
-    use sysinfo::System;
+fn collect_metrics(state: &AppState) -> Result<Metrics, String> {
+    let cpu_count = state.cpu_count();
+    let cpu_used_pct_rounded = state.cpu_used_pct();
 
-    let mut sys = System::new();
+    let mut sys = sysinfo::System::new();
     sys.refresh_memory();
-    sys.refresh_cpu_all();
-
-    // sysinfo needs a small delay for accurate CPU — first call returns 0.
-    // In a real daemon this would be cached; for now, report instantaneous.
-    std::thread::sleep(std::time::Duration::from_millis(100));
-    sys.refresh_cpu_all();
-
-    let cpu_count = sys.cpus().len() as u32;
-    let cpu_used_pct = sys.global_cpu_usage();
-    let cpu_used_pct_rounded = if cpu_used_pct > 0.0 {
-        (cpu_used_pct * 100.0).round() / 100.0
-    } else {
-        0.0
-    };
-
     let mem_total = sys.total_memory();
     let mem_used = sys.used_memory();
     let mem_total_mib = mem_total / 1024 / 1024;