From 56c04b1f70b7a4ee6b6b0a4646da4a1a6b0064ca Mon Sep 17 00:00:00 2001
From: Alexander Heldt <me@alexanderheldt.se>
Date: Sun, 26 Jul 2026 07:32:16 +0000
Subject: [PATCH] manatee: Guard podman containers against crash-looping

Add a start-limit (5 restarts / 5 min) to the romm, romm-db, romm-redis
and homeassistant podman services so a failing container enters a
`failed` state instead of retrying forever. A slow crash-loop (~5s per
attempt) stays under systemd's default 5-starts-per-10s limit, so a
longer window is needed to catch it.
---
 hosts/manatee/modules/home-assistant/default.nix | 10 ++++++++++
 hosts/manatee/modules/romm/default.nix           | 14 ++++++++++++++
 2 files changed, 24 insertions(+)

diff --git a/hosts/manatee/modules/home-assistant/default.nix b/hosts/manatee/modules/home-assistant/default.nix
index 458a3a7..bb870c0 100644
--- a/hosts/manatee/modules/home-assistant/default.nix
+++ b/hosts/manatee/modules/home-assistant/default.nix
@@ -117,6 +117,16 @@ in
 
     hardware.bluetooth.enable = true;
 
+    # Give up and enter a `failed` state (visible in `systemctl --failed`) if the
+    # container restarts more than 5 times in 5 minutes, instead of crash-looping
+    # forever. The window is deliberately long: a slow crash-loop (~5s per attempt)
+    # stays under systemd's default 5-starts-per-10s limit and would otherwise
+    # retry indefinitely.
+    systemd.services.podman-homeassistant = {
+      startLimitIntervalSec = 300;
+      startLimitBurst = 5;
+    };
+
     virtualisation.oci-containers = {
       backend = "podman";
 
diff --git a/hosts/manatee/modules/romm/default.nix b/hosts/manatee/modules/romm/default.nix
index 14d1008..ae6a61c 100644
--- a/hosts/manatee/modules/romm/default.nix
+++ b/hosts/manatee/modules/romm/default.nix
@@ -26,6 +26,16 @@ let
             - 'Thumbs.db'
             - '.DS_Store'
   '';
+
+  # Give up and enter a `failed` state (visible in `systemctl --failed`) if a
+  # container restarts more than 5 times in 5 minutes, instead of crash-looping
+  # forever. The window is deliberately long: a slow crash-loop (~5s per attempt)
+  # stays under systemd's default 5-starts-per-10s limit and would otherwise
+  # retry indefinitely.
+  crashLoopGuard = {
+    startLimitIntervalSec = 300;
+    startLimitBurst = 5;
+  };
 in
 {
   options = {
@@ -51,6 +61,10 @@ in
       "d /var/lib/romm/assets 0755 root root -"
     ];
 
+    systemd.services.podman-romm = crashLoopGuard;
+    systemd.services.podman-romm-db = crashLoopGuard;
+    systemd.services.podman-romm-redis = crashLoopGuard;
+
     systemd.services.romm-net = {
       description = "Create Podman network for RomM";
       after = [ "podman.service" ];