fix(stability): стадия 1 — таймауты на все TG RPC + устойчивый воркер загрузок + супервизия #8
+12
-4
@@ -1083,10 +1083,18 @@ async def ping() -> JSONResponse:
|
||||
# age is None right after boot: the watchdog hasn't run its first probe yet. Treat that
|
||||
# as healthy (gate on connected only) so a freshly-started container is not killed before
|
||||
# its first probe cycle — otherwise start_period would have to cover a full watchdog interval.
|
||||
# Note: when the watchdog is DISABLED (TG_WATCHDOG_ENABLED=false) age stays None forever,
|
||||
# so /ping degenerates to a pure connectivity check and cannot detect a stale-but-connected
|
||||
# ("zombie") session — that TG-liveness signal only exists while the watchdog runs.
|
||||
healthy = connected and (age is None or age < threshold)
|
||||
#
|
||||
# The staleness branch (age >= threshold => degraded) is only meaningful while the watchdog
|
||||
# is running to refresh age. With the watchdog DISABLED (TG_WATCHDOG_ENABLED=false) nothing
|
||||
# refreshes age — yet a disconnect-flap restart can still stamp it once (see _restart_client,
|
||||
# which runs before the watchdog-enabled gate), after which age only grows. Letting that
|
||||
# stale age drive /ping to 503 would spuriously fail the container healthcheck on a live
|
||||
# connection and trigger an autoheal restart. So gate staleness on the watchdog being on;
|
||||
# with it off, /ping is a pure connectivity check (no zombie-session detection — that
|
||||
# TG-liveness signal only exists while the watchdog runs).
|
||||
healthy = connected and (
|
||||
not Config["tg_watchdog_enabled"] or age is None or age < threshold
|
||||
)
|
||||
return JSONResponse(
|
||||
{
|
||||
"status": "ok" if healthy else "degraded",
|
||||
|
||||
@@ -99,6 +99,21 @@ def test_ping_degraded_stale_probe(patch_client):
|
||||
assert fake.client.get_me_calls == 0
|
||||
|
||||
|
||||
def test_ping_watchdog_disabled_stale_age_still_healthy(patch_client, monkeypatch):
|
||||
# With the watchdog OFF, nothing refreshes age (a disconnect-flap restart can stamp it
|
||||
# once, then it only grows). A stale age must NOT drive /ping to 503 on a live connection
|
||||
# — that would spuriously fail the healthcheck and trigger an autoheal restart. So with the
|
||||
# watchdog disabled, /ping is a pure connectivity check: connected + stale age => healthy.
|
||||
monkeypatch.setitem(api_server.Config, "tg_watchdog_enabled", False)
|
||||
fake, tc = patch_client(age=THRESHOLD + 100, is_connected=True)
|
||||
r = tc.get("/ping")
|
||||
assert r.status_code == 200
|
||||
body = r.json()
|
||||
assert body["status"] == "ok"
|
||||
assert body["connected"] is True
|
||||
assert fake.client.get_me_calls == 0
|
||||
|
||||
|
||||
def test_ping_degraded_disconnected(patch_client):
|
||||
# Even with a fresh probe age, a disconnected client is unhealthy.
|
||||
fake, tc = patch_client(age=1.0, is_connected=False)
|
||||
|
||||
Reference in New Issue
Block a user