bvandeusenandClaude Opus 5.5
a9bf4b7f16
After Steward or its database was down, every agent's age and every series' last sample were old, so all agent-offline rules fired and all rules went no-data at once, clearing a minute later. With per-account alert email that would have mailed everyone a dozen notices per restart. A periodic job now ticks a Watch (steward/core/watch.py) once it has reached the database; the Watch says since when runs have been unbroken (process start, or the end of a gap over 120 s). seconds_since_seen and the no-data sweep count silence from no earlier than that, so a source must stay quiet while Steward is watching before it counts as silent. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
53 lines
1.9 KiB
Python
53 lines
1.9 KiB
Python
"""Pure-function tests for the host_agent seconds_since_seen reconciler."""
|
|
from dataclasses import dataclass
|
|
from datetime import datetime, timedelta, timezone
|
|
|
|
from plugins.host_agent.scheduler import seconds_since_seen
|
|
|
|
|
|
@dataclass
|
|
class FakeReg:
|
|
host_id: str
|
|
last_seen_at: datetime | None
|
|
created_at: datetime | None
|
|
|
|
|
|
NOW = datetime(2026, 4, 15, 12, 0, 0, tzinfo=timezone.utc)
|
|
NAMES = {"h1": "nas", "h2": "web", "h3": "db"}
|
|
|
|
|
|
def test_seconds_since_each_agent_last_reported():
|
|
regs = [FakeReg("h1", NOW - timedelta(seconds=30), NOW - timedelta(days=9)),
|
|
FakeReg("h2", NOW - timedelta(minutes=10), NOW - timedelta(days=9))]
|
|
assert seconds_since_seen(regs, NAMES, NOW) == [("nas", 30.0), ("web", 600.0)]
|
|
|
|
|
|
def test_an_agent_that_never_reported_counts_from_registration():
|
|
regs = [FakeReg("h3", None, NOW - timedelta(minutes=5))]
|
|
assert seconds_since_seen(regs, NAMES, NOW) == [("db", 300.0)]
|
|
|
|
|
|
def test_a_registration_without_its_host_is_skipped():
|
|
regs = [FakeReg("gone", NOW, NOW)]
|
|
assert seconds_since_seen(regs, NAMES, NOW) == []
|
|
|
|
|
|
def test_a_clock_ahead_of_ours_reads_as_just_seen():
|
|
regs = [FakeReg("h1", NOW + timedelta(seconds=5), NOW)]
|
|
assert seconds_since_seen(regs, NAMES, NOW) == [("nas", 0.0)]
|
|
|
|
|
|
def test_empty_input():
|
|
assert seconds_since_seen([], NAMES, NOW) == []
|
|
|
|
|
|
def test_steward_s_own_outage_is_not_the_agent_s_silence():
|
|
"""#5441: agents that last reported 10 minutes ago, while Steward has only
|
|
been back for a minute, are a minute old; an agent heard since then keeps
|
|
its own age."""
|
|
back = NOW - timedelta(minutes=1)
|
|
regs = [FakeReg("h1", NOW - timedelta(minutes=10), NOW - timedelta(days=9)),
|
|
FakeReg("h2", NOW - timedelta(seconds=20), NOW - timedelta(days=9)),
|
|
FakeReg("h3", None, NOW - timedelta(days=9))]
|
|
assert seconds_since_seen(regs, NAMES, NOW, back) == [("nas", 60.0), ("web", 20.0), ("db", 60.0)]
|