fix: correct zero-safe pathconf checks and connectivity prefix match

- Use `is not None` for pathconf values so 0 is not silently dropped
- Broaden connectivity prefix check to catch bare "connectivity" key

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

.claude/settings.json

@@ -1,20 +0,0 @@
-{
-  "permissions": {
-    "allow": [
-      "Edit(*)",
-      "Bash(pytest *)",
-      "Bash(python *)",
-      "Bash(python3 *)",
-      "Bash(.venv/bin/pytest *)",
-      "Bash(npm *)",
-      "Bash(git *)",
-      "Bash(ls *)",
-      "Bash(cat *)",
-      "Bash(grep *)",
-      "Bash(find *)",
-      "Bash(mkdir *)",
-      "Bash(touch *)",
-      "Bash(uv *)"
-    ]
-  }
-}

.gitea/workflows/release.yml

@@ -33,7 +33,7 @@ jobs:
       - name: Generate changelog
         id: changelog
         run: |
-          PREV_TAG=$(git tag --sort=-version:refname | grep -v "^${GITHUB_REF#refs/tags/}$" | head -1)
+          PREV_TAG=$(git tag --sort=-version:refname | grep -m 1 -v "^${GITHUB_REF#refs/tags/}$")
           if [ -n "$PREV_TAG" ]; then
             CHANGELOG=$(git log --pretty=format:"- %s" "${PREV_TAG}..HEAD")
           else

.gitignore

@@ -5,6 +5,7 @@ __pycache__/
 *.pyo
 .flake8
 .venv/
+.continue/
 test/
 build/
 dist/

CHANGELOG.md

@@ -0,0 +1,457 @@
+# Changelog
+
+All notable changes to this project are documented here, organized by release.
+
+## [5.3.10]
+
+### Added
+- clear stale plugin data and persist OAuth users to config
+- auto-scale CPU history graph Y axis
+- add CPU usage history graph to CPU Monitor section
+
+### Fixed
+- remove bak file in bumpminor.sh
+
+---
+
+## [5.3.9]
+
+### Added
+- auto-update CHANGELOG and README in bumpminor.sh
+
+---
+
+## [5.3.8]
+
+### Added
+- Wiki home page with overview and getting started guide
+
+### Fixed
+- Release workflow: use `GITHUB_REF`/`GITHUB_OUTPUT` (Gitea Actions uses GitHub-compatible variable names)
+- Release workflow: replace `head -1` with `grep -m 1` to avoid SIGPIPE (exit 141) in changelog step
+
+---
+
+## [5.3.7]
+
+### Added
+- Dark mode with light/dark/auto theme setting
+- UNKNOWN level filter in Log of Events
+- Per-metric grace period input in threshold settings
+- Replace Dynamic DNS YAML editor with a web form
+- Sort hosts, thresholds, and channels alphabetically on settings page
+- Suppress alerts for unwatched hosts
+
+### Fixed
+- Preserve log message order when replaying history on connect
+
+---
+
+## [5.3.6]
+
+### Added
+- MIT license
+
+### Fixed
+- Correct ZFS pool status threshold operator and add per-metric grace
+- Normalize email and domain fields
+- Move dependencies back under `[project]` in pyproject.toml
+
+---
+
+## [5.3.4]
+
+### Fixed
+- Run full reload after HTTP config publish, not just `config.reload()`
+
+---
+
+## [5.3.3]
+
+### Added
+- Replace YAML threshold editor with a form-based UI
+- Replace multi-select fields with dual-panel picker on settings page
+- Nav bar button to publish pending config changes
+- Host, level, and message filters in Log of Events
+
+### Fixed
+- Remove container max-width; stop stretching inputs on settings page
+
+### Removed
+- Legacy `dyndnshosts`/`drophosts` config keys
+
+---
+
+## [5.3.2]
+
+### Added
+- Retry DNS resolution indefinitely; add `-4`/`-6` address-family flags to `hbc` and `hbc_mini`
+- Replace YAML hosts editor with form-based CRUD table
+- Replace YAML notification channel editor with form-based UI
+
+### Fixed
+- Support list-valued `threshold_config` in hosts table
+- Derive hosts threshold config list from config file keys
+- Replace channel checkboxes in Users table with multi-select
+- Support plugin-level `enabled: false` in threshold config
+- Always populate glance strip for all hosts on page load
+- Fetch host info on initial page load
+
+---
+
+## [5.3.1]
+
+### Added
+- Host info section in Host Overview (fetched and rendered on card expand)
+- `GET /api/0/hosts/{hostname}/info` endpoint
+- Show suffix-matched metric coverage in host info threshold table
+- Move `hbc_version` and `hbc_type` out of `os_info` into the host info section
+
+### Fixed
+- Correct `THRESHOLD_DEFAULTS` metric keys and add missing defaults
+
+---
+
+## [5.3.0]
+
+### Added
+- Profile page self-service: change identity, password, and notification channels
+- Settings page editor with form sections, YAML editors, stage/publish/rollback workflow
+- Config read API: `GET /api/0/config`, `/section/{name}`, `/backups`
+- Config write API: `POST /api/0/config`, `POST /api/0/config/rollback`
+- `configio` module for comment-preserving YAML round-trip writes
+- Multi-provider OAuth2 login page and generic provider routes
+- Log login/logout events to the event log with auth source
+
+### Fixed
+- ZFS monitor alerts dropped on restart with wildcard pool thresholds
+- Preserve OAuth users across config reload
+- Config API error handling, consistent 403 messages, deduplicated key lists
+- Validate password body type; coerce `notification_channels` to strings in profile API
+- Preserve OAuth `client_secret` on roundtrip; harden rollback path validation
+
+---
+
+## [5.2.6]
+
+### Added
+- Alerts host-filter field with URL query parameter and notify URL
+- Optional logo on Gitea OAuth login button
+
+### Fixed
+- Show human-readable duration in re-notification messages
+
+---
+
+## [5.2.5]
+
+### Added
+- Alert CRITICAL on degraded or suspended ZFS pools (ONLINE=OK, DEGRADED=WARNING, all else=CRITICAL)
+- Sign in with Gitea button on login page with OAuth2 redirect/callback routes
+- OAuth2 CSRF state management
+- Host owner shown in glance strip for admin users
+- C port of `hbc_mini` (single-file client in `scripts/c/`)
+
+### Fixed
+- Use `base_url` config for OAuth redirect URI to handle reverse proxy deployments
+- Preserve OAuth users across config reload
+- Escape HTML in login page error display
+
+---
+
+## [5.2.4]
+
+### Added
+- `hbc`/`hbc_mini`: `owner` config field included in `os_info`; server applies to host record
+- Server requests InfoPlugin refresh when a host has no plugin data
+- Event log stores structured dicts; filter by user
+
+### Fixed
+- Strip `_status_code` suffix from displayed metric names in threshold alerts
+- Use plain URL in Mattermost plugin metrics link
+- Fall back to `default_owner` when `os_info` has no owner
+
+---
+
+## [5.2.3]
+
+### Added
+- `hbc`/`hbc_mini`: log name and version at startup
+- Show metric name inline with hostname in alerts and notifications
+
+### Fixed
+- Send shutdown message only if a boot message was previously sent; suppress both on restart
+
+---
+
+## [5.2.2]
+
+### Fixed
+- Retry connection on network error instead of permanently dropping it
+- Silence `aiohttp.access` log; strip plugin prefix in alerts UI
+
+---
+
+## [5.2.1]
+
+### Fixed
+- Threshold and logging improvements
+
+---
+
+## [5.2.0]
+
+### Added
+- `nagios` operator for direct exit-code severity mapping
+
+### Fixed
+- Always show `THRESHOLD_DEFAULTS` in Settings threshold config
+
+---
+
+## [5.1.21]
+
+### Added
+- `nagios_runner` improvements and alerts page fixes
+
+---
+
+## [5.1.20]
+
+### Added
+- Generic threshold matching for `nagios_runner` with `{check_name}` display support
+
+### Fixed
+- Reduce default hysteresis from 10% to 2%
+- Show recovery threshold in alerts UI
+
+---
+
+## [5.1.19]
+
+### Added
+- Exclude ZFS ARC from `memory_percent`
+- Add `uptime_seconds` to `cpu_monitor`
+
+### Fixed
+- Send boot/shutdown message on the first open connection, not blindly on the first in list
+
+---
+
+## [5.1.18]
+
+### Added
+- Fetch-based Update/Delete buttons with toast notifications on Host Overview
+
+### Fixed
+- Settings thresholds show correct per-config metrics; miscellaneous `hbc` fixes
+
+---
+
+## [5.1.17]
+
+### Added
+- Owner Update/Delete buttons on Host Overview; purge stale alerts on reload
+- Retry `AsyncConnection.open()` indefinitely; drop IPv6 only on early startup failure
+- Alert pie chart in the nav bar
+
+### Fixed
+- Make Alerts page scrollable
+
+---
+
+## [5.1.16]
+
+### Added
+- Generic `ping_monitor` thresholds; round RTT to nearest ms
+
+---
+
+## [5.1.15]
+
+### Added
+- Link hostnames in Live Dashboard to Host Overview
+- Threshold Configurations section on settings page
+
+### Fixed
+- Suppress notifications on alert de-escalation (e.g. CRITICAL→WARNING)
+- Suppress recover messages for down durations under 4 seconds
+
+---
+
+## [5.1.14]
+
+### Added
+- ZFS pool renderer in Host Overview
+
+---
+
+## [5.1.13]
+
+### Added
+- ZFS monitor plugin
+- Host-level watch flag to suppress notifications
+- Filter Live Dashboard and Host Overview by owner/manager
+- Composable `threshold_config` list for per-host threshold layering
+- Restart on SIGHUP in `hbc` and `hbc_mini`
+
+### Fixed
+- Mask `api_password` and `access_token` in settings page
+
+---
+
+## [5.1.12]
+
+Internal release — no user-visible changes.
+
+---
+
+## [5.1.11]
+
+### Fixed
+- Install under Docker
+- Clean up install script
+
+---
+
+## [5.1.10]
+
+### Fixed
+- Synchronize version in `hbc_mini`
+- Install script no longer overwrites itself
+
+---
+
+## [5.1.9]
+
+### Added
+- Install `hbc_mini` via package or install script
+
+---
+
+## [5.1.8]
+
+### Added
+- Track `hbc` type and version
+
+### Fixed
+- Nav bar position
+
+---
+
+## [5.1.7]
+
+### Added
+- `hbc_mini`: single-file heartbeat client
+
+### Fixed
+- Drop dead connections on protocol error
+
+---
+
+## [5.1.6]
+
+### Fixed
+- Simplify event log usage; fix argument handling
+
+---
+
+## [5.1.5]
+
+### Added
+- Update `hbc` via `hb_install.sh` instead of code patching
+
+---
+
+## [5.1.4]
+
+### Added
+- Redesign Plugin Metrics page as Host Overview
+
+---
+
+## [5.1.3]
+
+### Added
+- Validate absolute command paths at `nagios_runner` init
+- Async subprocess in `nagios_runner` with stderr capture and signal handling
+- `skip_reason` field on `Plugin`; surface in `PluginLoader` init messaging
+
+### Fixed
+- Use `shlex.split()` for `nagios_runner` path validation to handle quoted paths
+- Reconfigure logging to syslog after `daemonize()`
+
+---
+
+## [5.1.2]
+
+### Fixed
+- Plugin config lookup shadowed by `CLIENT_DEFAULTS` plugins key
+- Apply grace period to all threshold alerts before logging/notifying
+- RECOVER routing: use consistent level name and route via alerted channel
+- Early reminder notifications and lost recovery notifications
+- Non-alerting of overdue hosts
+
+### Added
+- Swiss clock widget in the UI
+
+---
+
+## [5.1.1]
+
+### Added
+- SMS and Matrix notification channels
+- CLI commands `stop`, `restart`, and `reload` for `hbd`
+- WebSocket endpoint at `http://.../ws`
+- Mobile HTML pages
+
+### Fixed
+- Profile not updating
+- Sortable columns in tables
+
+---
+
+## [5.1.0]
+
+### Added
+- Ping monitor plugin
+- Persist state to pickle file; restart timers on server restart
+- SIGHUP config reload for `hbd`
+- Renotify on CRITICAL only; persistent user sessions
+- RTT count threshold
+
+### Fixed
+- Bogus notification on new clients
+- Show "overdue" in alerts instead of null
+
+---
+
+## [5.0.12]
+
+### Added
+- User management and settings page
+
+---
+
+## [5.0.10]
+
+### Added
+- Publish package to Gitea PyPI registry
+
+---
+
+## [5.0.9]
+
+### Added
+- Use `SO_TIMESTAMP` for RTT measurement (Linux, FreeBSD, macOS)
+- Persist state to pickle file; restart timers on restart
+
+---
+
+## [5.0.6]
+
+### Added
+- Major codebase refactoring: restructured into client/server components
+- Per-client threshold configuration
+- Display and acknowledge alerts in the UI
+- Proper `hbc` termination; `hbd` config reloadable at runtime

README.md

@@ -20,7 +20,7 @@ A lightweight UDP-based host monitoring system. Monitored hosts run a client (`h
   └────────────────────┘              └────────────────────────────┘
 ```
 
-**Package:** `hbd` v5.3.4
+**Package:** `hbd` v5.3.10
 **Python:** 3.11+
 
 ### Subpackages

hbd/__init__.py

@@ -14,4 +14,4 @@ Install options:
 """
 
 __all__ = ["__version__"]
-__version__ = "5.3.7"
+__version__ = "5.3.10"

hbd/client/plugins/filesystem_info.py

@@ -127,7 +127,7 @@ class FilesystemInfoPlugin(InfoPlugin):
                     try:
                         # Maximum filename length
                         max_name = os.pathconf(partition.mountpoint, 'PC_NAME_MAX')
-                        if max_name:
+                        if max_name is not None:
                             fs_info['maxfile'] = max_name
                     except (OSError, ValueError):
                         pass
@@ -135,7 +135,7 @@ class FilesystemInfoPlugin(InfoPlugin):
                     try:
                         # Maximum path length
                         max_path = os.pathconf(partition.mountpoint, 'PC_PATH_MAX')
-                        if max_path:
+                        if max_path is not None:
                             fs_info['maxpath'] = max_path
                     except (OSError, ValueError):
                         pass

hbd/server/hbdclass.py

@@ -297,6 +297,8 @@ class Host:
         self.plugin_retention = 100  # Keep last N samples per plugin
         # Alert state tracking: {metric_path: AlertState}
         self.alert_states = {}
+        # Stale-data timers: {plugin_name: asyncio.TimerHandle}
+        self.plugin_timers = {}
         # User access control
         self.owner: str | None = None       # username of owner
         self.managers: list = []            # usernames with manager role
@@ -365,7 +367,7 @@ class Host:
     def stateinfo(self):
         ddict = {}
         for d in self.__dict__:
-            if d in ["alert_states", "plugin_data"]:
+            if d in ["alert_states", "plugin_data", "plugin_timers"]:
                 continue
             if d == "connections":
                 cl = []
@@ -483,6 +485,8 @@ class Host:
             self.managers = []
         if not hasattr(self, "monitors"):
             self.monitors = []
+        if not hasattr(self, "plugin_timers"):
+            self.plugin_timers = {}
 
         pass
 
@@ -542,6 +546,34 @@ class Host:
         """
         return self.plugin_data
 
+    def reset_plugin_timer(self, plugin_name, timeout_seconds, callback):
+        """Reset the stale-data timer for a plugin.
+
+        If no new PLG data arrives within timeout_seconds, callback(host, plugin_name)
+        is called so the caller can clear history and alerts.
+        """
+        import asyncio
+        existing = self.plugin_timers.get(plugin_name)
+        if existing and not existing.cancelled():
+            existing.cancel()
+
+        async def _fire():
+            await callback(self, plugin_name)
+
+        try:
+            loop = asyncio.get_event_loop()
+            self.plugin_timers[plugin_name] = loop.call_later(
+                timeout_seconds, lambda: asyncio.create_task(_fire())
+            )
+        except RuntimeError:
+            pass
+
+    def cancel_plugin_timer(self, plugin_name):
+        """Cancel the stale timer for a plugin, if any."""
+        handle = self.plugin_timers.pop(plugin_name, None)
+        if handle and not handle.cancelled():
+            handle.cancel()
+
     # ------------------------------------------------------------------
     # User-role helpers
     # ------------------------------------------------------------------

hbd/server/http.py

@@ -424,7 +424,7 @@ async def start(
         # Resolve templates directory relative to the hbd package
         pkg_dir = os.path.dirname(__file__)
         templates_dir = config.get("templates_dir", os.path.join(pkg_dir, "templates"))
-        env = jinja2.Environment(loader=jinja2.FileSystemLoader(templates_dir))
+        env = jinja2.Environment(loader=jinja2.FileSystemLoader(templates_dir), autoescape=True)
         host = config.get("hb_host", "localhost")
         extra_scripts = config.get("http_extra_scripts", "")
         host = request.host  # includes port if non-standard
@@ -597,8 +597,6 @@ async def start(
         all_alerts = []
 
         for hostname, host in hbdclass.Host.hosts.items():
-            if not host.watched:
-                continue
             if not _can_view_host(user, host):
                 continue
             if threshold_checker:
@@ -692,7 +690,7 @@ async def start(
         current_user, _ = _require_auth_redirect(request)
         pkg_dir = os.path.dirname(__file__)
         templates_dir = config.get("templates_dir", os.path.join(pkg_dir, "templates"))
-        env = jinja2.Environment(loader=jinja2.FileSystemLoader(templates_dir))
+        env = jinja2.Environment(loader=jinja2.FileSystemLoader(templates_dir), autoescape=True)
 
         # Collect all hosts with plugin data (filtered by visibility)
         hosts_with_plugins = []
@@ -723,7 +721,7 @@ async def start(
         current_user, _ = _require_auth_redirect(request)
         pkg_dir = os.path.dirname(__file__)
         templates_dir = config.get("templates_dir", os.path.join(pkg_dir, "templates"))
-        env = jinja2.Environment(loader=jinja2.FileSystemLoader(templates_dir))
+        env = jinja2.Environment(loader=jinja2.FileSystemLoader(templates_dir), autoescape=True)
 
         tmpl = env.get_template("alerts.html")
         body = tmpl.render(
@@ -780,6 +778,8 @@ async def start(
                 token = users_mod.create_session(username)
                 eventlog("hbd", "INFO", f"Login: {username} via password")
                 redirect_to = request.rel_url.query.get("next", "/")
+                if not redirect_to.startswith("/"):
+                    redirect_to = "/"
                 resp = web.HTTPFound(redirect_to)
                 resp.set_cookie(
                     SESSION_COOKIE,
@@ -891,6 +891,13 @@ async def start(
         if not target_user.avatar_is_local():
             return web.Response(status=404, text="No local avatar configured")
         path = target_user.avatar
+        avatar_dir = config.get("avatar_dir") or (
+            os.path.dirname(os.path.realpath(_config_path)) if _config_path else None
+        )
+        if not avatar_dir:
+            return web.Response(status=403, text="Local avatars not configured")
+        if not os.path.realpath(path).startswith(os.path.realpath(avatar_dir) + os.sep):
+            return web.Response(status=403, text="Forbidden")
         if not os.path.isfile(path):
             return web.Response(status=404, text="Avatar file not found")
         # Infer content-type from extension
@@ -994,7 +1001,7 @@ async def start(
         current_user, _ = _require_auth_redirect(request)
         pkg_dir = os.path.dirname(__file__)
         templates_dir = config.get("templates_dir", os.path.join(pkg_dir, "templates"))
-        env = jinja2.Environment(loader=jinja2.FileSystemLoader(templates_dir))
+        env = jinja2.Environment(loader=jinja2.FileSystemLoader(templates_dir), autoescape=True)
 
         # Build host access summary for this user.
         # Merge live hosts with config-only hosts (not yet seen) so the profile
@@ -1078,7 +1085,7 @@ async def start(
         current_user, _ = _require_auth_redirect(request)
         pkg_dir = os.path.dirname(__file__)
         templates_dir = config.get("templates_dir", os.path.join(pkg_dir, "templates"))
-        env = jinja2.Environment(loader=jinja2.FileSystemLoader(templates_dir))
+        env = jinja2.Environment(loader=jinja2.FileSystemLoader(templates_dir), autoescape=True)
         from hbd import __version__ as hbd_version
 
         uptime_secs = int(time.time() - _start_epoch)
@@ -1122,7 +1129,7 @@ async def start(
             raise web.HTTPForbidden(reason="Admin access required")
         pkg_dir = os.path.dirname(__file__)
         templates_dir = config.get("templates_dir", os.path.join(pkg_dir, "templates"))
-        env = jinja2.Environment(loader=jinja2.FileSystemLoader(templates_dir))
+        env = jinja2.Environment(loader=jinja2.FileSystemLoader(templates_dir), autoescape=True)
         tmpl = env.get_template("settings.html")
         settings_data = settings_mod.get_settings_data(config, threshold_checker=threshold_checker)
         body = tmpl.render(
@@ -1182,6 +1189,23 @@ async def start(
             profile["full_name"],
             profile["avatar_url"],
         )
+        # Persist new OAuth users to the config file so they survive restarts.
+        # Only write when the user isn't already in the config's users section.
+        if _config_path and not (config.get("users") or {}).get(user.username):
+            try:
+                disk_data = configio_mod.read_roundtrip(_config_path)
+                if not disk_data.get("users"):
+                    disk_data["users"] = {}
+                disk_data["users"][user.username] = {
+                    k: v for k, v in [
+                        ("full_name", user.full_name),
+                        ("avatar", user.avatar),
+                    ] if v
+                }
+                configio_mod.write_config(_config_path, disk_data)
+                logger.info("Persisted OAuth user %r to config", user.username)
+            except Exception as exc:
+                logger.warning("Failed to persist OAuth user %r to config: %s", user.username, exc)
         session_token = users_mod.create_session(user.username)
         eventlog("hbd", "INFO", f"Login: {user.username} via {provider.type}")
         resp = web.HTTPFound("/")
@@ -1644,7 +1668,16 @@ async def start(
             if "full_name" in body:
                 user_entry["full_name"] = str(body["full_name"])
             if "avatar" in body:
-                user_entry["avatar"] = str(body["avatar"])
+                avatar_val = str(body["avatar"])
+                if avatar_val.startswith("/"):
+                    avatar_dir = config.get("avatar_dir") or (
+                        os.path.dirname(os.path.realpath(_config_path)) if _config_path else None
+                    )
+                    if not avatar_dir:
+                        return web.json_response({"error": "Local avatars not configured"}, status=400)
+                    if not os.path.realpath(avatar_val).startswith(os.path.realpath(avatar_dir) + os.sep):
+                        return web.json_response({"error": "Avatar path outside allowed directory"}, status=400)
+                user_entry["avatar"] = avatar_val
             if "notification_channels" in body:
                 visible = _visible_channels_for_user(user)
                 user_entry["notification_channels"] = [

hbd/server/notify.py

@@ -140,7 +140,9 @@ def _send_pushover(channel_cfg: dict, notif: Notification) -> bool:
     if not token or not user:
         logger.warning("pushover: missing token or user")
         return False
-    params: dict = {"token": token, "user": user, "title": notif.title, "message": notif.body}
+    body = "%s: %s" % (notif.title, notif.body)
+    title = ""
+    params: dict = {"token": token, "user": user, "title": title, "message": body}
     if channel_cfg.get("sound"):
         params["sound"] = channel_cfg["sound"]
     if notif.url:

hbd/server/templates/live.html

@@ -321,9 +321,15 @@
     var c = 0;
     var HBD_VERSION = "{{ hbd_version }}";
 
+    function escHtml(s) {
+      var d = document.createElement('div');
+      d.textContent = String(s);
+      return d.innerHTML;
+    }
+
     function hostNameHtml(data) {
       var rawName = data.raw_name || data.name.replace(/<[^>]+>/g, '').replace('*', '').trim();
-      var nameHtml = data.name;
+      var nameHtml = escHtml(data.name);
       if (!data.hbc_version || data.hbc_version !== HBD_VERSION) {
         nameHtml += ' 🥀';
       }
@@ -410,11 +416,11 @@
         c_critical.innerHTML = "";
       }
       
-      c_ipv4addr.innerHTML = data.connections[0].addr;
+      c_ipv4addr.innerHTML = escHtml(data.connections[0].addr);
-      c_ipv4state.innerHTML = data.connections[0].state;
+      c_ipv4state.innerHTML = escHtml(data.connections[0].state);
       if (data.connections.length > 1) {
-        c_ipv6addr.innerHTML = data.connections[1].addr;
+        c_ipv6addr.innerHTML = escHtml(data.connections[1].addr);
-        c_ipv6state.innerHTML = data.connections[1].state;
+        c_ipv6state.innerHTML = escHtml(data.connections[1].state);
       }
       var table = document.getElementById("ntablebody"); // find table to append to
       table.appendChild(row); // append row to table
@@ -477,7 +483,7 @@
 
       for  (var i = 0; i < data.connections.length; i++) {
         // Offset by 2 for the warning/critical count columns
-        name_idx[data.name].cells[3 + i * 4].innerHTML = data.connections[i].addr;
+        name_idx[data.name].cells[3 + i * 4].innerHTML = escHtml(data.connections[i].addr);
         name_idx[data.name].cells[6 + i * 4].innerHTML = formatTS(
           data.connections[i].statetime
         );
@@ -497,7 +503,7 @@
             state = '<span class="state-overdue">overdue</span>';
             latency = "-";
           } else {
-            state = "<b>" + data.connections[i].state + "</b>";
+            state = "<b>" + escHtml(data.connections[i].state) + "</b>";
             latency = "-";
           }
         }
@@ -558,12 +564,12 @@
                        + ' ' + _p(_d.getHours()) + ':' + _p(_d.getMinutes()) + ':' + _p(_d.getSeconds());
             var lvl = (msg.level || "INFO").toLowerCase();
             var hostVal = msg.host || '';
-            var html = '<div class="log-entry log-' + lvl + '" data-level="' + lvl + '" data-host="' + hostVal.replace(/"/g, '&quot;') + '">';
+            var html = '<div class="log-entry log-' + escHtml(lvl) + '" data-level="' + escHtml(lvl) + '" data-host="' + escHtml(hostVal) + '">';
             html += '<span class="log-ts">' + ts_str + '</span>';
-            html += '<span class="log-level">' + (msg.level || "") + '</span>';
+            html += '<span class="log-level">' + escHtml(msg.level || "") + '</span>';
-            if (msg.host) html += '<span class="log-host">' + msg.host + '</span>';
+            if (msg.host) html += '<span class="log-host">' + escHtml(msg.host) + '</span>';
-            if (msg.service) html += '<span class="log-service">' + msg.service + '</span>';
+            if (msg.service) html += '<span class="log-service">' + escHtml(msg.service) + '</span>';
-            html += '<span class="log-msg">' + msg.message + '</span>';
+            html += '<span class="log-msg">' + escHtml(msg.message) + '</span>';
             html += '</div>';
             msgs.insertAdjacentHTML(state.history ? "beforeend" : "afterbegin", html);
             applyLogFilters();
@@ -621,7 +627,7 @@
           <tbody id="ntablebody">
             {% for host in hosts %}
             <tr class="{% if host.alert_critical_unacked > 0 or host.alert_critical_acked > 0 %}row-critical{% elif host.alert_warning_unacked > 0 or host.alert_warning_acked > 0 %}row-warning{% endif %}">
-              <td data-name="{{ host.name }}"><a class="host-link" href="/plugins#{{ host.raw_name | urlencode }}">{{ host.name }}{% if not host.hbc_version or host.hbc_version != hbd_version %} 🥀{% endif %}</a></td>
+              <td data-name="{{ host.name }}"><a class="host-link" href="/plugins#{{ host.name | urlencode }}">{{ host.name }}{% if not host.hbc_version or host.hbc_version != hbd_version %} 🥀{% endif %}</a></td>
               <td style="text-align: center; color: #ff9800; font-weight: bold;">
                 {%- set warning_unacked = host.alert_warning_unacked -%}
                 {%- set warning_acked = host.alert_warning_acked -%}

hbd/server/templates/plugins.html

@@ -914,7 +914,7 @@
         let html = '';
         switch (pluginName) {
           case 'os_info':        html = renderOsInfoTable(cached.data); break;
-          case 'cpu_monitor':    html = renderCpuTable(cached.data); break;
+          case 'cpu_monitor':    html = renderCpuTable(hostname, cached.data); break;
           case 'memory_monitor': html = renderMemoryTable(cached.data); break;
           case 'disk_monitor':   html = renderDiskTables(cached.data); break;
           case 'network_monitor':html = renderNetworkTables(cached.data); break;
@@ -926,6 +926,10 @@
 
         html += `<div class="timestamp">Last updated: ${new Date(cached.timestamp * 1000).toLocaleString()}</div>`;
         body.innerHTML = html;
+
+        if (pluginName === 'cpu_monitor') {
+          fetchCpuHistory(hostname).then(samples => renderCpuChart(hostname, samples)).catch(() => {});
+        }
       }
 
       // ── Per-plugin renderers ────────────────────────────────────────────────
@@ -948,7 +952,92 @@
         return html;
       }
 
-      function renderCpuTable(d) {
+      async function fetchCpuHistory(hostname) {
+        const r = await fetch(`/api/0/hosts/${encodeURIComponent(hostname)}/plugins/cpu_monitor?limit=100`);
+        if (!r.ok) return [];
+        const json = await r.json();
+        return json.samples || [];
+      }
+
+      function renderCpuChart(hostname, samples) {
+        const el = document.getElementById(`cpu-chart-${hostname}`);
+        if (!el || !samples.length) return;
+
+        const pts = samples
+          .filter(s => s.data.cpu_percent != null)
+          .map(s => ({ t: s.timestamp, v: s.data.cpu_percent }));
+        if (pts.length < 2) { el.style.display = 'none'; return; }
+
+        const W = 600, H = 80, PAD = { top: 6, right: 8, bottom: 18, left: 28 };
+        const cW = W - PAD.left - PAD.right;
+        const cH = H - PAD.top - PAD.bottom;
+
+        const tMin = pts[0].t, tMax = pts[pts.length - 1].t;
+        const tRange = tMax - tMin || 1;
+        const x = t => PAD.left + ((t - tMin) / tRange) * cW;
+
+        // Auto-scale Y axis with 10% padding, clamped to [0, 100]
+        const vMin = Math.min(...pts.map(p => p.v));
+        const vMax = Math.max(...pts.map(p => p.v));
+        const vRange = vMax - vMin || 1;
+        const vPad = Math.max(vRange * 0.1, 1);
+        const yLow  = Math.max(0,   vMin - vPad);
+        const yHigh = Math.min(100, vMax + vPad);
+        const yRange = yHigh - yLow || 1;
+        const y = v => PAD.top + cH - ((v - yLow) / yRange) * cH;
+
+        // Build polyline points and filled area path
+        const linePoints = pts.map(p => `${x(p.t).toFixed(1)},${y(p.v).toFixed(1)}`).join(' ');
+        const areaPath = `M${x(pts[0].t).toFixed(1)},${(PAD.top + cH).toFixed(1)} ` +
+          pts.map(p => `L${x(p.t).toFixed(1)},${y(p.v).toFixed(1)}`).join(' ') +
+          ` L${x(pts[pts.length-1].t).toFixed(1)},${(PAD.top + cH).toFixed(1)} Z`;
+
+        // Color based on latest absolute CPU %
+        const latest = pts[pts.length - 1].v;
+        const strokeColor = latest > 90 ? '#e53935' : latest > 70 ? '#fb8c00' : '#43a047';
+        const fillColor   = latest > 90 ? '#ffcdd2' : latest > 70 ? '#ffe0b2' : '#c8e6c9';
+
+        // Compute nice tick step for ~3-5 grid lines
+        const rawStep = yRange / 4;
+        const mag = Math.pow(10, Math.floor(Math.log10(rawStep || 1)));
+        const niceStep = [1, 2, 5, 10].map(f => f * mag).find(s => yRange / s <= 5) || mag * 10;
+        const tickStart = Math.ceil(yLow / niceStep) * niceStep;
+        let gridLines = '';
+        for (let v = tickStart; v <= yHigh + 0.001; v += niceStep) {
+          const yy = y(v).toFixed(1);
+          const label = Number.isInteger(v) ? v : v.toFixed(1);
+          gridLines += `<line x1="${PAD.left}" y1="${yy}" x2="${PAD.left + cW}" y2="${yy}" stroke="#e0e0e0" stroke-width="1"/>`;
+          gridLines += `<text x="${(PAD.left - 3).toFixed(1)}" y="${yy}" text-anchor="end" dominant-baseline="middle" font-size="8" fill="#999">${label}</text>`;
+        }
+
+        // X-axis time labels
+        const fmt = ts => {
+          const d = new Date(ts * 1000);
+          return d.toLocaleTimeString([], { hour: '2-digit', minute: '2-digit' });
+        };
+        const xLabels = `
+          <text x="${PAD.left}" y="${H - 2}" text-anchor="start" font-size="8" fill="#999">${fmt(pts[0].t)}</text>
+          <text x="${PAD.left + cW}" y="${H - 2}" text-anchor="end" font-size="8" fill="#999">${fmt(pts[pts.length-1].t)}</text>`;
+
+        el.innerHTML = `<svg viewBox="0 0 ${W} ${H}" preserveAspectRatio="none"
+          style="width:100%;height:${H}px;display:block;">
+          <defs>
+            <clipPath id="cpu-clip-${hostname}">
+              <rect x="${PAD.left}" y="${PAD.top}" width="${cW}" height="${cH}"/>
+            </clipPath>
+          </defs>
+          ${gridLines}
+          <line x1="${PAD.left}" y1="${PAD.top}" x2="${PAD.left}" y2="${PAD.top + cH}" stroke="#ccc" stroke-width="1"/>
+          <line x1="${PAD.left}" y1="${PAD.top + cH}" x2="${PAD.left + cW}" y2="${PAD.top + cH}" stroke="#ccc" stroke-width="1"/>
+          <g clip-path="url(#cpu-clip-${hostname})">
+            <path d="${areaPath}" fill="${fillColor}" opacity="0.6"/>
+            <polyline points="${linePoints}" fill="none" stroke="${strokeColor}" stroke-width="1.5" stroke-linejoin="round"/>
+          </g>
+          ${xLabels}
+        </svg>`;
+      }
+
+      function renderCpuTable(hostname, d) {
         const KEYS = [
           ['cpu_percent',      'CPU Usage',         'bar'],
           ['load_1min',        'Load (1 min)',       'num'],
@@ -966,7 +1055,8 @@
         ];
 
         const handled = new Set(KEYS.map(r => r[0]));
-        let html = '<table class="data-table"><thead><tr><th>Metric</th><th>Value</th></tr></thead><tbody>';
+        let html = `<div id="cpu-chart-${hostname}" style="margin-bottom:8px;"></div>`;
+        html += '<table class="data-table"><thead><tr><th>Metric</th><th>Value</th></tr></thead><tbody>';
         for (const [k, label, fmt] of KEYS) {
           if (!(k in d)) continue;
           const v = d[k];

hbd/server/threshold.py

@@ -1554,6 +1554,10 @@ class ThresholdChecker:
             configured = self.get_thresholds_for_host(hostname)
             stale = []
             for mp in host.alert_states:
+                # connectivity.* and rtt are managed by the connection state
+                # machine, not by threshold config — never purge them.
+                if mp == "rtt" or mp.startswith("connectivity"):
+                    continue
                 if self._find_threshold(configured, mp)[0] is not None:
                     continue
                 # Also match wildcard pool/partition thresholds (e.g. "zfs_monitor.*.status"

hbd/server/udp.py

@@ -232,6 +232,23 @@ def _make_timer_callbacks(uname, host, ctx):
     return on_overdue, on_unknown
 
 
+def _make_plugin_stale_callback(uname, ctx):
+    """Return an async callback that clears stale plugin data and its alerts."""
+    msg_to_websockets = ctx.get("msg_to_websockets")
+
+    async def on_plugin_stale(host, plugin_name):
+        host.plugin_data.pop(plugin_name, None)
+        stale_keys = [k for k in host.alert_states if k.startswith(f"{plugin_name}.")]
+        for k in stale_keys:
+            del host.alert_states[k]
+        eventlog(uname, "INFO", f"plugin data stale: {plugin_name}")
+        if msg_to_websockets:
+            msg_to_websockets("plugin_stale", {"host": uname, "plugin": plugin_name})
+            msg_to_websockets("host", host.stateinfo())
+
+    return on_plugin_stale
+
+
 def restore_connection_timers(hbdclass, ctx):
     """Restore overdue timers for all loaded connections after a pickle restore.
 
@@ -249,10 +266,15 @@ def restore_connection_timers(hbdclass, ctx):
         for afam, conn in list(host.connections.items()):
             state = conn.getstate()
             if state == hbdclass.Connection.DOWN:
+                _set_connectivity_alert(host, afam, "CRITICAL")
                 continue
 
             on_overdue, on_unknown = _make_timer_callbacks(uname, host, ctx)
 
+            if state == hbdclass.Connection.UNKNOWN:
+                _set_connectivity_alert(host, afam, "CRITICAL")
+                continue
+
             if state == hbdclass.Connection.UP and interval > 0:
                 elapsed = now - conn.lastbeat
                 # Give hosts one full (interval + grace) of extra time on startup
@@ -283,6 +305,10 @@ def restore_connection_timers(hbdclass, ctx):
                         "Restored OVERDUE timer %s/%s: %.0fs remaining",
                         uname, afam, remaining,
                     )
+                # Ensure the connectivity alert is set — it may be missing if
+                # hbd was shut down before the on_overdue callback had a chance
+                # to record it.
+                _set_connectivity_alert(host, afam, "CRITICAL")
                 restored += 1
 
     logger.info("Restored timers for %d connection(s)", restored)
@@ -372,12 +398,33 @@ def handle_datagram(msg: dict, addr, transport, ctx: dict):
                            if k not in ("ID", "plugin", "id", "name")}
             # Store plugin data with timestamp
             host.add_plugin_data(plugin_name, plugin_data, timestamp=now)
+            # Reset stale timer using the observed send interval for this plugin.
+            # We need two samples to know the real interval; on the first sample
+            # we cancel any leftover timer but don't set a new one, to avoid
+            # false-stale firing for slow plugins (e.g. nagios_runner at 300 s).
+            history = host.plugin_data.get(plugin_name, [])
+            if len(history) >= 2:
+                plugin_interval = max(history[-1][0] - history[-2][0], 1)
+                host.reset_plugin_timer(plugin_name, plugin_interval * 3,
+                                        _make_plugin_stale_callback(uname, ctx))
+                # Remove alert states for metrics present in the previous sample
+                # but absent now (e.g. a nagios check removed from configuration).
+                prev_keys = set(history[-2][1].keys())
+                curr_keys = set(plugin_data.keys())
+                for metric_name in prev_keys - curr_keys:
+                    metric_path = f"{plugin_name}.{metric_name}"
+                    if host.alert_states.pop(metric_path, None) is not None:
+                        eventlog(uname, "INFO", f"stale check removed: {metric_path}")
+                if (prev_keys - curr_keys) and msg_to_websockets:
+                    msg_to_websockets("host", host.stateinfo())
+            else:
+                host.cancel_plugin_timer(plugin_name)
 
             # If os_info reports an owner and none is configured server-side, apply it
             if plugin_name == "os_info":
                 config_owner =  config_mod.get_host_access(cfg, uname).get("owner")
                 default_owner = config_mod.get_default_owner(cfg)
-                inferred_owner = plugin_data.get("owner", config_owner or default_owner)    
+                inferred_owner = config_owner or plugin_data.get("owner") or default_owner
                 host.owner = inferred_owner
                 logger.info(f"owner for {uname} is {host.owner}")
             if DEBUG > 1:
@@ -432,6 +479,7 @@ def handle_datagram(msg: dict, addr, transport, ctx: dict):
     boot = msg.get("boot", 0)
 
     if boot:
+        # hbc was stared with a -b flag
         eventlog(uname, "INFO", "booted")
         if host.watched:
             asyncio.create_task(notify_mod.send_notification(
@@ -439,11 +487,24 @@ def handle_datagram(msg: dict, addr, transport, ctx: dict):
                 notify_mod.Notification(title=f"[INFO] {uname}", body=f"{host.name} booted", level="INFO"),
             ))
     if message:
-        eventlog(uname, "INFO", "msg: %s" % message, service=service)
+        eventlog(uname, "INFO", message, service=service)
 
     if conn.getstate() != hbdcls.Connection.UP:
+        # Transition to UP and log/notify if appropriate
         lasts = conn.state
         d = conn.newstate(hbdcls.Connection.UP, now)
+        # On reboot, pre-boot plugin data and derived alerts are stale.
+        # Cancel all plugin timers and wipe plugin state so timers restart
+        # cleanly from the first two post-boot samples.
+        for pname in list(host.plugin_timers):
+            host.cancel_plugin_timer(pname)
+        host.plugin_data.clear()
+        stale_plugin_keys = [
+            k for k in host.alert_states
+            if k not in ("rtt",) and not k.startswith("connectivity.")
+        ]
+        for k in stale_plugin_keys:
+            del host.alert_states[k]
         # Clear connectivity alert now that the host is back up
         _set_connectivity_alert(host, conn.afam, "OK")
         # Don't log/notify RECOVER for a brand-new host seen for the first time —

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "hbd"
-version = "5.3.7"
+version = "5.3.10"
 description = "Heartbeat monitoring system — client (hbc) and server (hbd)"
 readme = "README.md"
 requires-python = ">=3.11"

scripts/bumpminor.sh

@@ -5,9 +5,23 @@ uv version --bump patch
 VER=$(uv  version  --short)
 sed -i".bak"  "s/__version__ = \"[0-9.]*\"\(.*\)$/__version__ = \"$VER\"\1/" hbd/__init__.py
 sed -i".bak"  "s/__version__ = \"[0-9.]*\"\(.*\)$/__version__ = \"$VER\"\1/" scripts/hbc_mini.py
+sed -i".bak"  "s/\*\*Package:\*\* \`hbd\` v[0-9.]*/\*\*Package:\*\* \`hbd\` v$VER/" README.md
+
+# Update CHANGELOG.md with commits since last tag
+LASTTAG=$(git describe --tags --abbrev=0 2>/dev/null || true)
+ADDED=$(git log "${LASTTAG:+$LASTTAG..}HEAD" --pretty="%s" | grep "^feat:" | sed 's/^feat: /- /')
+FIXED=$(git log "${LASTTAG:+$LASTTAG..}HEAD" --pretty="%s" | grep "^fix:" | sed 's/^fix: /- /')
+{
+    printf "## [%s]\n" "$VER"
+    [ -n "$ADDED" ] && printf "\n### Added\n%s\n" "$ADDED"
+    [ -n "$FIXED" ] && printf "\n### Fixed\n%s\n" "$FIXED"
+    printf "\n---\n\n"
+} > /tmp/changelog_entry.txt
+sed -i".bak" "4r /tmp/changelog_entry.txt" CHANGELOG.md
+rm /tmp/changelog_entry.txt CHANGELOG.md.bak
 
 # commit pyproject.toml
-git commit -m "version $VER" pyproject.toml hbd/__init__.py scripts/hbc_mini.py
+git commit -m "version $VER" pyproject.toml hbd/__init__.py scripts/hbc_mini.py README.md CHANGELOG.md
 git push 
 # tag version
 git tag -a v$VER -m "Version $VER"
@@ -15,3 +29,4 @@ git push --tags
 
 rm hbd/__init__.py.bak
 rm scripts/hbc_mini.py.bak
+rm README.md.bak

scripts/c/hbc_mini.c

@@ -789,7 +789,7 @@ static void plugin_cpu_monitor(conn_t *c, const config_t *cfg) {
  * Plugin: memory_monitor
  * Linux:   /proc/meminfo
  * FreeBSD: sysctl vm.stats.vm.*
- * NetBSD:  sysctl vm.uvmexp (struct uvmexp)
+ * NetBSD:  sysctl vm.uvmexp (struct uvmexp_sysctl)
  * ============================================================ */
 
 /* emit the common kvdict fields and send */
@@ -896,9 +896,9 @@ static void plugin_memory_monitor(conn_t *c, const config_t *cfg) {
 
 static void plugin_memory_monitor(conn_t *c, const config_t *cfg) {
     (void)cfg;
-    struct uvmexp uvm;
+    struct uvmexp_sysctl uvm;
     size_t len = sizeof(uvm);
-    int mib[2] = {CTL_VM, VM_UVMEXP};
+    int mib[2] = {CTL_VM, VM_UVMEXP2};
     if (sysctl(mib, 2, &uvm, &len, NULL, 0) != 0) return;
 
     long long ps   = uvm.pagesize;

scripts/hbc_mini.py

@@ -41,7 +41,7 @@ from pathlib import Path
 from typing import Any, Dict, List, Optional, Tuple
 
 # updated by scripts/bumpminor.sh
-__version__ = "5.3.7"
+__version__ = "5.3.10"
 
 # ---------------------------------------------------------------------------
 # Protocol  (mirrors hbd/common/proto.py)

scripts/hbc_windows.spec

@@ -0,0 +1,49 @@
+# PyInstaller spec for hbc_windows.exe
+# Build with: pyinstaller hbc_windows.spec
+#
+# Requirements (on Windows):
+#   pip install pyinstaller
+
+block_cipher = None
+
+a = Analysis(
+    ['hbc_windows.py'],
+    pathex=[],
+    binaries=[],
+    datas=[],
+    hiddenimports=[],
+    hookspath=[],
+    hooksconfig={},
+    runtime_hooks=[],
+    excludes=['tkinter', 'unittest', 'email', 'html', 'http', 'urllib', 'xml'],
+    win_no_prefer_redirects=False,
+    win_private_assemblies=False,
+    cipher=block_cipher,
+    noarchive=False,
+)
+
+pyz = PYZ(a.pure, a.zlib_archive, cipher=block_cipher)
+
+exe = EXE(
+    pyz,
+    a.scripts,
+    a.binaries,
+    a.zipfiles,
+    a.datas,
+    [],
+    name='hbc_windows',
+    debug=False,
+    bootloader_ignore_signals=False,
+    strip=False,
+    upx=False,
+    upx_exclude=[],
+    runtime_tmpdir=None,
+    console=True,
+    disable_windowed_traceback=False,
+    argv_emulation=False,
+    target_arch=None,
+    codesign_identity=None,
+    entitlements_file=None,
+    icon=None,
+    version=None,
+)

scripts/install_hbc_windows.ps1

@@ -0,0 +1,126 @@
+#Requires -RunAsAdministrator
+<#
+.SYNOPSIS
+    Install hbc_windows.exe as a Windows Service using NSSM.
+
+.DESCRIPTION
+    Installs the HeartBeat Client as a Windows Service that starts automatically.
+    Requires NSSM (Non-Sucking Service Manager) in PATH or alongside this script.
+    Requires hbc_windows.exe built via: pyinstaller hbc_windows.spec
+
+.PARAMETER Server
+    HBD server hostname or IP address (required).
+
+.PARAMETER ExePath
+    Path to hbc_windows.exe. Defaults to the directory containing this script.
+
+.PARAMETER ServiceName
+    Windows service name. Default: heartbeat-client
+
+.PARAMETER ConfigFile
+    Path to hbc.json config file. Optional.
+
+.PARAMETER LogFile
+    Path to log file. Default: C:\ProgramData\heartbeat\hbc.log
+
+.PARAMETER Interval
+    Heartbeat interval in seconds. Default: 10
+
+.EXAMPLE
+    .\install_hbc_windows.ps1 -Server hbd.example.com
+    .\install_hbc_windows.ps1 -Server hbd.example.com -ConfigFile C:\ProgramData\heartbeat\hbc.json
+#>
+
+param(
+    [Parameter(Mandatory = $true)]
+    [string]$Server,
+
+    [string]$ExePath = "",
+    [string]$ServiceName = "heartbeat-client",
+    [string]$ConfigFile = "",
+    [string]$LogFile = "C:\ProgramData\heartbeat\hbc.log",
+    [int]$Interval = 10
+)
+
+Set-StrictMode -Version Latest
+$ErrorActionPreference = "Stop"
+
+# Locate hbc_windows.exe
+if ($ExePath -eq "") {
+    $ExePath = Join-Path $PSScriptRoot "hbc_windows.exe"
+}
+if (-not (Test-Path $ExePath)) {
+    Write-Error "hbc_windows.exe not found at: $ExePath`nBuild it first with: pyinstaller hbc_windows.spec"
+    exit 1
+}
+
+# Locate NSSM
+$nssm = Get-Command nssm -ErrorAction SilentlyContinue
+if (-not $nssm) {
+    $nssmLocal = Join-Path $PSScriptRoot "nssm.exe"
+    if (Test-Path $nssmLocal) {
+        $nssm = $nssmLocal
+    } else {
+        Write-Error "nssm.exe not found in PATH or alongside this script.`nDownload from https://nssm.cc/download"
+        exit 1
+    }
+} else {
+    $nssm = $nssm.Source
+}
+
+# Build argument list
+$args_list = "--daemon $Server"
+if ($ConfigFile -ne "") {
+    $args_list = "--daemon -c `"$ConfigFile`" $Server"
+}
+if ($LogFile -ne "") {
+    $args_list = "$args_list --log-file `"$LogFile`""
+}
+
+# Create data directory
+$dataDir = "C:\ProgramData\heartbeat"
+if (-not (Test-Path $dataDir)) {
+    New-Item -ItemType Directory -Path $dataDir | Out-Null
+    Write-Host "Created $dataDir"
+}
+
+# Remove existing service if present
+$existing = Get-Service -Name $ServiceName -ErrorAction SilentlyContinue
+if ($existing) {
+    Write-Host "Removing existing service '$ServiceName'..."
+    & $nssm stop $ServiceName 2>$null
+    & $nssm remove $ServiceName confirm
+}
+
+# Install service
+Write-Host "Installing service '$ServiceName'..."
+& $nssm install $ServiceName $ExePath $args_list
+if ($LASTEXITCODE -ne 0) {
+    Write-Error "nssm install failed (exit $LASTEXITCODE)"
+    exit 1
+}
+
+# Configure service
+& $nssm set $ServiceName DisplayName "HeartBeat Client"
+& $nssm set $ServiceName Description "Sends heartbeat and plugin metrics to the HBD monitoring server."
+& $nssm set $ServiceName Start SERVICE_AUTO_START
+& $nssm set $ServiceName AppStdout (Join-Path $dataDir "nssm_stdout.log")
+& $nssm set $ServiceName AppStderr (Join-Path $dataDir "nssm_stderr.log")
+& $nssm set $ServiceName AppRotateFiles 1
+& $nssm set $ServiceName AppRotateBytes 5242880
+
+# Start service
+Write-Host "Starting service '$ServiceName'..."
+& $nssm start $ServiceName
+if ($LASTEXITCODE -ne 0) {
+    Write-Warning "Service installed but failed to start — check logs in $dataDir"
+} else {
+    Write-Host "Service '$ServiceName' started successfully."
+    Write-Host "Log file: $LogFile"
+    Write-Host ""
+    Write-Host "Useful commands:"
+    Write-Host "  nssm status $ServiceName"
+    Write-Host "  nssm stop $ServiceName"
+    Write-Host "  nssm restart $ServiceName"
+    Write-Host "  nssm remove $ServiceName confirm"
+}