hermes-agent/tests/gateway/test_loop_liveness_watchdog.py
Teknium 39975613b1
test: prune wave 2 + speed fixes — 28,106 → 19,757 test functions, suite wall 315s → 294s
Second, deeper pass over tools/gateway/hermes_cli plus first pass over
the trees wave 1 missed (acp, acp_adapter, skills, computer_use, docker,
dashboard, conformance, monitoring, secret_sources, hermes_state,
providers). Same rubric as wave 1 (AGENTS.md test policy); security,
alternation/caching invariants, issue-number regressions, and E2E kept.

Real test-quality fixes found and rooted out along the way:
- tests/tools/test_command_guards.py made real auxiliary-LLM HTTPS calls
  (DEFAULT_CONFIG smart-approval leaked in) — pinned approval
  mode=manual via autouse fixture: 17.4s → 0.4s.
- test_model_switch_custom_providers.py / test_user_providers_model_switch.py
  silently probed live provider catalogs (~2s/test) — stubbed
  cached_provider_model_ids/provider_model_ids/fetch_api_models.
- test_telegram_noise_filter.py: 15-platform copy-paste matrix over
  shared gateway.run logic → 3 representative platforms (55s → 3.9s).
- test_gateway_shutdown.py: stop()'s 5s interrupt-deadline loop spun on
  MagicMock agents — interrupt.side_effect now clears _running_agents
  (22s → 1.0s).
- test_gateway_inactivity_timeout.py poll-harness timings shrunk 3-5x
  (24s → 1.1s); test_mcp_stability.py backoff/SIGTERM-grace sleeps
  patched (15.4s → 2.5s); test_async_delegation.py negative-drain wait
  5s → 0.5s.
- test_telegram_init_deadline.py: loop-block margin restored to 1.0s
  with rationale comment — the watchdog-dump assertion needs the loop
  blocked well past deadline+grace under parallel load (flaked once in
  the 40-worker verification run at a 0.2s margin).

Verification: full hermetic suite via scripts/run_tests.sh —
2,438 files, 21,718 tests passed, 0 failed, 293.9s wall.
Suite totals vs original baseline: 46,820 → 19,757 test functions
(−57.8%), wall 583.5s → 293.9s (−50%), subprocess CPU 13,564s → 11,623s.
2026-07-29 13:39:40 -07:00

200 lines
6.6 KiB
Python

"""Gateway event-loop freeze backstops for issue #69089."""
from __future__ import annotations
import asyncio
import threading
import time
from unittest.mock import MagicMock, patch
import pytest
from gateway.shutdown_watchdog import (
_arm_loop_floor_timer,
start_loop_liveness_watchdog,
)
def _immediate_loop() -> MagicMock:
loop = MagicMock(spec=asyncio.AbstractEventLoop)
loop.call_soon_threadsafe.side_effect = lambda callback: callback()
return loop
def test_loop_liveness_watchdog_stop_during_dump_disarms_hard_exit():
loop = MagicMock(spec=asyncio.AbstractEventLoop)
handle_ready = threading.Event()
handle_ref = {}
exit_codes = []
def stop_during_dump(*_args, **_kwargs) -> None:
assert handle_ready.wait(timeout=2.0)
handle_ref["handle"].stop()
with (
patch("gateway.shutdown_watchdog.logger.critical") as critical,
patch(
"gateway.shutdown_watchdog.faulthandler.dump_traceback",
side_effect=stop_during_dump,
) as dump,
patch("gateway.shutdown_watchdog.os._exit", side_effect=exit_codes.append),
):
handle = start_loop_liveness_watchdog(
loop, probe_interval=0.01, probe_timeout=0.01, max_strikes=1
)
assert handle is not None
handle_ref["handle"] = handle
handle_ready.set()
handle.join(timeout=2.0)
assert not handle.is_alive()
critical.assert_called_once()
dump.assert_called_once_with(all_threads=True)
assert exit_codes == []
def test_loop_liveness_watchdog_stop_during_final_miss_disarms_hard_exit():
loop = MagicMock(spec=asyncio.AbstractEventLoop)
probe_scheduled = threading.Event()
release_probe = threading.Event()
probe_event_ref = {}
handle_ref = {}
exit_codes = []
class FinalStrikeLimit:
def __gt__(self, _strikes: int) -> bool:
# If strike evaluation is reached, keep recheck #2 from masking a
# missing post-probe recheck #1 in this boundary test.
handle_ref["handle"]._stop_event.clear()
return False
def hold_scheduled_probe(callback) -> None:
probe_event_ref["event"] = callback.__self__
probe_scheduled.set()
assert release_probe.wait(timeout=2.0)
loop.call_soon_threadsafe.side_effect = hold_scheduled_probe
with (
patch("gateway.shutdown_watchdog.logger.critical") as critical,
patch("gateway.shutdown_watchdog.faulthandler.dump_traceback") as dump,
patch("gateway.shutdown_watchdog.os._exit", side_effect=exit_codes.append),
):
handle = start_loop_liveness_watchdog(
loop,
probe_interval=0.01,
probe_timeout=0.01,
max_strikes=FinalStrikeLimit(),
)
assert handle is not None
handle_ref["handle"] = handle
assert probe_scheduled.wait(timeout=2.0), "watchdog did not schedule a probe"
def stop_during_miss() -> bool:
handle.stop()
return False
probe_event_ref["event"].is_set = stop_during_miss
release_probe.set()
handle.join(timeout=1.0)
assert not handle.is_alive()
assert exit_codes == []
critical.assert_not_called()
dump.assert_not_called()
def test_loop_liveness_watchdog_stop_after_first_recheck_skips_final_actions():
loop = MagicMock(spec=asyncio.AbstractEventLoop)
probe_scheduled = threading.Event()
release_probe = threading.Event()
def hold_scheduled_probe(callback) -> None:
probe_scheduled.set()
assert release_probe.wait(timeout=2.0)
loop.call_soon_threadsafe.side_effect = hold_scheduled_probe
with (
patch("gateway.shutdown_watchdog.logger.critical") as critical,
patch("gateway.shutdown_watchdog.faulthandler.dump_traceback") as dump,
patch("gateway.shutdown_watchdog.os._exit") as hard_exit,
):
handle = start_loop_liveness_watchdog(
loop, probe_interval=0.01, probe_timeout=0.01, max_strikes=1
)
assert handle is not None
assert probe_scheduled.wait(timeout=2.0), "watchdog did not schedule a probe"
original_is_set = handle._stop_event.is_set
is_set_calls = 0
def stop_on_final_recheck() -> bool:
nonlocal is_set_calls
is_set_calls += 1
# With the forced immediate timeout: _wait_for_probe is call 1,
# recheck #1 is call 2, and recheck #2 is call 3.
if is_set_calls == 3:
handle.stop()
return original_is_set()
handle._stop_event.is_set = stop_on_final_recheck
with patch(
"gateway.shutdown_watchdog.time.monotonic", side_effect=[0.0, 1.0]
):
release_probe.set()
handle.join(timeout=1.0)
assert is_set_calls == 3
assert not handle.is_alive()
critical.assert_not_called()
dump.assert_not_called()
hard_exit.assert_not_called()
def test_gateway_config_loop_watchdog_round_trip():
"""loop_watchdog is a config.yaml knob: default on, nested-gateway form honored."""
from gateway.config import GatewayConfig
assert GatewayConfig.from_dict({}).loop_watchdog is True
assert GatewayConfig.from_dict({"loop_watchdog": False}).loop_watchdog is False
assert (
GatewayConfig.from_dict(
{"gateway": {"loop_watchdog": "off"}}
).loop_watchdog
is False
)
config = GatewayConfig.from_dict({"loop_watchdog": False})
assert config.to_dict()["loop_watchdog"] is False
def test_gateway_runner_liveness_guards_start_and_stop():
from gateway.run import GatewayRunner
runner = object.__new__(GatewayRunner)
runner._loop_floor_timer_handle = None
runner._loop_liveness_watchdog = None
loop = MagicMock(spec=asyncio.AbstractEventLoop)
floor_timer = MagicMock()
watchdog = MagicMock()
watchdog.is_alive.return_value = True
with (
patch(
"gateway.run._arm_loop_floor_timer", return_value=floor_timer
) as arm_floor,
patch(
"gateway.run.start_loop_liveness_watchdog", return_value=watchdog
) as start_watchdog,
):
runner._start_loop_liveness_guards(loop)
arm_floor.assert_called_once_with(loop)
start_watchdog.assert_called_once_with(loop)
assert runner._loop_floor_timer_handle is floor_timer
assert runner._loop_liveness_watchdog is watchdog
runner._stop_loop_liveness_guards()
watchdog.stop.assert_called_once_with()
floor_timer.cancel.assert_called_once_with()
assert runner._loop_liveness_watchdog is None
assert runner._loop_floor_timer_handle is None