fix: clear flap state when a host is dropped

flap._state persists at module level keyed by (host, service) and only
clears via a RECOVER-triggered quiet window. A host dropped mid-flap with
no RECOVER ever received leaves ok_since permanently None, so the
flapping flag could never clear on its own.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

hbd/server/flap.py

@@ -100,6 +100,17 @@ def observe(host: str, service: str, level: str) -> str:
     return SUPPRESS
 
 
+def clear_host(host: str) -> None:
+    """Discard all flap state for *host* (called when a host is dropped).
+
+    A dropped host may be mid-flap with no RECOVER ever received, in which
+    case ``ok_since`` stays ``None`` and ``_sweep`` can never clear it on its
+    own — the state would otherwise persist forever.
+    """
+    for key in [k for k in _state if k[0] == host]:
+        del _state[key]
+
+
 def flapping_services(host: str) -> list:
     """Return the services of *host* that are currently flapping.
 

hbd/server/http.py

@@ -14,6 +14,7 @@ import logging
 from aiohttp import web
 import jinja2
 from . import data
+from . import flap as flap_mod
 from . import notify as notify_mod
 from . import settings as settings_mod
 from . import users as users_mod
@@ -439,6 +440,7 @@ async def start(
             return web.json_response({"error": "Forbidden"}, status=403)
         eventlog(uname, "INFO", "dropped")
         del hbdclass.Host.hosts[uname]
+        flap_mod.clear_host(uname)
         return web.Response(text="Done")
 
     async def register(request):

tests/test_flap.py

@@ -98,6 +98,26 @@ def test_alert_during_the_quiet_window_keeps_it_flapping():
     assert flap.flapping_services("h1") == ["cpu"]
 
 
+# --- host removal ------------------------------------------------------------
+
+def test_clear_host_drops_flapping_state_even_without_a_recover():
+    # A host dropped mid-flap (no RECOVER ever received) must not stay
+    # flapping forever: nothing will ever set ok_since for it again.
+    alerts(4)
+    assert flap.flapping_services("h1") == ["cpu"]
+    flap.clear_host("h1")
+    assert flap.flapping_services("h1") == []
+    assert flap.observe("h1", "cpu", "CRITICAL") == flap.PASS
+
+
+def test_clear_host_only_affects_the_named_host():
+    alerts(4, host="h1")
+    alerts(4, host="h2")
+    flap.clear_host("h1")
+    assert flap.flapping_services("h1") == []
+    assert flap.flapping_services("h2") == ["cpu"]
+
+
 # --- keying -----------------------------------------------------------------
 
 def test_services_and_hosts_are_tracked_independently():