From c7af43ab33669106fb7b4ffefa6881fca5cce346 Mon Sep 17 00:00:00 2001
From: Claude <noreply@anthropic.com>
Date: Fri, 27 Mar 2026 10:29:52 +0000
Subject: [PATCH] feat: restart stopped app containers on server startup

Adds restore_app_containers() which runs at startup alongside
restore_caddy_routes(). For each app with a successful deploy it
inspects the container state via `podman inspect` and runs
`podman start` if the container is exited (e.g. after a host reboot).
Missing containers are logged as warnings requiring a manual redeploy.

https://claude.ai/code/session_01FKCW3FDjNFj6jve4niMFXH
---
 server/src/main.rs        |  6 ++++
 server/src/routes/apps.rs | 67 +++++++++++++++++++++++++++++++++++++++
 2 files changed, 73 insertions(+)

diff --git a/server/src/main.rs b/server/src/main.rs
index 152bfe8..6bb8821 100644
--- a/server/src/main.rs
+++ b/server/src/main.rs
@@ -162,6 +162,12 @@ async fn main() -> anyhow::Result<()> {
         routes::apps::restore_caddy_routes(&restore_db).await;
     });
 
+    // Restart any app containers that are stopped (e.g. after a host reboot).
+    let containers_db = state.db.clone();
+    tokio::spawn(async move {
+        routes::apps::restore_app_containers(&containers_db).await;
+    });
+
     // ── Protected routes (admin login required) ───────────────────────────────
     let protected = Router::new()
         .route("/",           get(routes::ui::index))
diff --git a/server/src/routes/apps.rs b/server/src/routes/apps.rs
index 58f8345..869ca4f 100644
--- a/server/src/routes/apps.rs
+++ b/server/src/routes/apps.rs
@@ -76,6 +76,73 @@ pub async fn restore_caddy_routes(db: &crate::DbPool) {
     tracing::info!("restore_caddy_routes: registered {} app routes", apps.len());
 }
 
+/// On startup, ensure every app that had a successful deploy is actually running.
+/// If the host rebooted, containers will be in "exited" state — start them.
+/// If a container is missing entirely, log a warning (we don't rebuild automatically).
+pub async fn restore_app_containers(db: &crate::DbPool) {
+    let apps = match sqlx::query_as::<_, crate::models::App>("SELECT * FROM apps")
+        .fetch_all(db)
+        .await
+    {
+        Ok(a)  => a,
+        Err(e) => { tracing::error!("restore_app_containers: DB error: {}", e); return; }
+    };
+
+    for app in &apps {
+        // Only care about apps that have at least one successful deploy.
+        let has_deploy: bool = sqlx::query_scalar(
+            "SELECT COUNT(*) > 0 FROM deploys WHERE app_id = ? AND status = 'success'"
+        )
+        .bind(&app.id)
+        .fetch_one(db)
+        .await
+        .unwrap_or(false);
+
+        if !has_deploy {
+            continue;
+        }
+
+        let container = format!("hiy-{}", app.id);
+
+        // Check container state via `podman inspect`.
+        let inspect = tokio::process::Command::new("podman")
+            .args(["inspect", "--format", "{{.State.Status}}", &container])
+            .output()
+            .await;
+
+        match inspect {
+            Ok(out) if out.status.success() => {
+                let status = String::from_utf8_lossy(&out.stdout).trim().to_string();
+                if status == "running" {
+                    tracing::debug!("restore_app_containers: {} already running", container);
+                } else {
+                    tracing::info!("restore_app_containers: starting {} (was {})", container, status);
+                    let start = tokio::process::Command::new("podman")
+                        .args(["start", &container])
+                        .output()
+                        .await;
+                    match start {
+                        Ok(o) if o.status.success() =>
+                            tracing::info!("restore_app_containers: {} started", container),
+                        Ok(o) =>
+                            tracing::warn!("restore_app_containers: failed to start {}: {}",
+                                container, String::from_utf8_lossy(&o.stderr).trim()),
+                        Err(e) =>
+                            tracing::warn!("restore_app_containers: error starting {}: {}", container, e),
+                    }
+                }
+            }
+            _ => {
+                tracing::warn!(
+                    "restore_app_containers: container {} not found — redeploy needed",
+                    container
+                );
+            }
+        }
+    }
+    tracing::info!("restore_app_containers: done");
+}
+
 /// Push a visibility change to Caddy without requiring a full redeploy.
 /// Best-effort: logs a warning on failure but does not surface an error to the caller.
 async fn push_visibility_to_caddy(app_id: &str, port: i64, is_public: bool) {