Phase 7d: per-container MemoryMax + CPUQuota via systemd drop-in

hive-c0re/src/lifecycle.rs

@@ -25,6 +25,11 @@ const GIT_EMAIL: &str = "hive-c0re@hyperhive";
 const WEB_PORT_BASE: u16 = 8100;
 const WEB_PORT_RANGE: u16 = 900;
 
+/// Default resource caps applied to every managed container via a systemd
+/// drop-in under `/run/systemd/system/container@<NAME>.service.d/`.
+const DEFAULT_MEMORY_MAX: &str = "2G";
+const DEFAULT_CPU_QUOTA: &str = "50%";
+
 /// Returns the per-agent web UI port. Same hash on both sides — manager,
 /// dashboard, and agent harness all agree.
 #[must_use]
@@ -69,6 +74,8 @@ pub async fn spawn(
     let flake_ref = format!("{}#default", applied_dir.display());
     run(&["create", &container, "--flake", &flake_ref]).await?;
     set_nspawn_flags(&container, agent_dir)?;
+    set_resource_limits(&container)?;
+    systemd_daemon_reload().await?;
     run(&["start", &container]).await
 }
 
@@ -89,6 +96,8 @@ pub async fn rebuild(
     let container = container_name(name);
     let flake_ref = format!("{}#default", applied_dir.display());
     set_nspawn_flags(&container, agent_dir)?;
+    set_resource_limits(&container)?;
+    systemd_daemon_reload().await?;
     run(&["update", &container, "--flake", &flake_ref]).await?;
     // Restart so any nspawn-level changes (bind mounts, networking, etc.) apply.
     run(&["stop", &container]).await?;
@@ -280,6 +289,42 @@ async fn git_status(dir: &Path, args: &[&str]) -> Result<bool> {
     Ok(st.success())
 }
 
+/// Write a systemd drop-in for `container@<container>.service` that applies
+/// our default resource caps. Goes under `/run/systemd/system/...` so it's
+/// ephemeral (regenerated on every spawn / rebuild).
+fn set_resource_limits(container: &str) -> Result<()> {
+    let dir = format!("/run/systemd/system/container@{container}.service.d");
+    std::fs::create_dir_all(&dir).with_context(|| format!("create {dir}"))?;
+    let path = format!("{dir}/hyperhive-limits.conf");
+    let content = format!(
+        "[Service]\nMemoryMax={DEFAULT_MEMORY_MAX}\nCPUQuota={DEFAULT_CPU_QUOTA}\n",
+    );
+    std::fs::write(&path, content).with_context(|| format!("write {path}"))?;
+    tracing::info!(
+        %path,
+        memory_max = DEFAULT_MEMORY_MAX,
+        cpu_quota = DEFAULT_CPU_QUOTA,
+        "wrote resource limits drop-in"
+    );
+    Ok(())
+}
+
+async fn systemd_daemon_reload() -> Result<()> {
+    let out = Command::new("systemctl")
+        .arg("daemon-reload")
+        .output()
+        .await
+        .context("invoke systemctl daemon-reload")?;
+    if !out.status.success() {
+        bail!(
+            "systemctl daemon-reload failed ({}): {}",
+            out.status,
+            String::from_utf8_lossy(&out.stderr).trim()
+        );
+    }
+    Ok(())
+}
+
 /// Idempotently rewrite the `EXTRA_NSPAWN_FLAGS` line in
 /// `/etc/nixos-containers/<container>.conf`. The start script expands this
 /// variable unquoted into the `systemd-nspawn` command.