fix: validate complete-day monitoring coverage

This commit is contained in:
xavierk
2026-09-29 03:34:17 +05:30
parent d045e88043
commit e22b99442e
7 changed files with 130 additions and 65 deletions
+22 -31
View File
@@ -33,6 +33,11 @@ from itertools import pairwise
from typing import Any
from zoneinfo import ZoneInfo
from .monitoring_periods import (
as_utc_datetime,
interval_within_one_monitoring_period,
)
@dataclass(frozen=True)
class LocalDaySummary:
@@ -308,27 +313,13 @@ def persist_local_day(
return created
def _utc_datetime(value: str | datetime) -> datetime:
parsed = value if isinstance(value, datetime) else datetime.fromisoformat(value)
if parsed.tzinfo is None:
parsed = parsed.replace(tzinfo=timezone.utc)
return parsed.astimezone(timezone.utc)
def _same_monitoring_period(
conn: sqlite3.Connection,
start: datetime,
end: datetime,
) -> bool:
"""Return whether one recorded monitoring period contains the interval."""
for started_at, ended_at in conn.execute(
"SELECT started_at, ended_at FROM monitoring_periods"
):
period_start = _utc_datetime(started_at)
period_end = _utc_datetime(ended_at) if ended_at else None
if period_start <= start and (period_end is None or end <= period_end):
return True
return False
return interval_within_one_monitoring_period(conn, start, end)
def repair_legacy_local_day_evidence(conn: sqlite3.Connection) -> int:
@@ -365,13 +356,13 @@ def repair_legacy_local_day_evidence(conn: sqlite3.Connection) -> int:
segment_windows = {}
if {"monitoring_periods", "controller_segments"}.issubset(tables):
periods = [
(_utc_datetime(start), _utc_datetime(end) if end else None)
(as_utc_datetime(start), as_utc_datetime(end) if end else None)
for start, end in conn.execute(
"SELECT started_at, ended_at FROM monitoring_periods"
)
]
segment_boundaries = [
(segment_id, _utc_datetime(opened_at))
(segment_id, as_utc_datetime(opened_at))
for segment_id, opened_at in conn.execute(
"SELECT id, opened_at FROM controller_segments ORDER BY opened_at, id"
)
@@ -399,7 +390,7 @@ def repair_legacy_local_day_evidence(conn: sqlite3.Connection) -> int:
)
samples = [
{
"id": row[0], "ts": _utc_datetime(row[1]),
"id": row[0], "ts": as_utc_datetime(row[1]),
"bytes_written": row[2], "bytes_read": row[3],
"segment_id": row[4], "local_tz": row[5],
}
@@ -453,8 +444,8 @@ def repair_legacy_local_day_evidence(conn: sqlite3.Connection) -> int:
"id": row[0],
"local_date": row[1],
"tz_name": row[2],
"start": _utc_datetime(row[3]),
"end": _utc_datetime(row[4]),
"start": as_utc_datetime(row[3]),
"end": as_utc_datetime(row[4]),
"precision": row[6],
}
for row in recorded_days
@@ -482,8 +473,8 @@ def repair_legacy_local_day_evidence(conn: sqlite3.Connection) -> int:
local_day_id, local_date, tz_name, utc_start, utc_end,
prior_incomplete,
) = row
start = _utc_datetime(utc_start)
end = _utc_datetime(utc_end)
start = as_utc_datetime(utc_start)
end = as_utc_datetime(utc_end)
matching = matching_by_day[local_day_id]
if not matching:
@@ -595,8 +586,8 @@ def _record_reconstructed_shared_intervals(
day_rows = [
{
"id": row[0], "local_date": row[1], "tz_name": row[2],
"utc_start": _utc_datetime(row[3]),
"utc_end": _utc_datetime(row[4]),
"utc_start": as_utc_datetime(row[3]),
"utc_end": as_utc_datetime(row[4]),
"precision": row[6],
}
for row in recorded_days
@@ -685,7 +676,7 @@ def _coarse_local_day_activity(
full_hours: dict[datetime, tuple] = {}
boundary_hour_found = False
other_day_bounds = [
(_utc_datetime(row[0]), _utc_datetime(row[1]))
(as_utc_datetime(row[0]), as_utc_datetime(row[1]))
for row in conn.execute(
"SELECT utc_start, utc_end FROM local_days WHERE id != ?",
(local_day_id,),
@@ -698,7 +689,7 @@ def _coarse_local_day_activity(
"FROM hour_observations WHERE hour >= ? AND hour < ? ORDER BY hour",
(first_hour.isoformat(), day_end.isoformat()),
):
hour_start = _utc_datetime(row[0])
hour_start = as_utc_datetime(row[0])
hour_end = hour_start + hour
if hour_start >= day_end or hour_end <= day_start:
continue
@@ -856,8 +847,8 @@ def record_local_activity_interval(
both samples share one monitoring period, and the interval stays inside one
recorded local date. Every other valid difference remains one unallocated row.
"""
start = _utc_datetime(previous["ts"])
end = _utc_datetime(current["ts"])
start = as_utc_datetime(previous["ts"])
end = as_utc_datetime(current["ts"])
if end <= start:
return None
@@ -1004,12 +995,12 @@ def mark_local_activity_gap(
previous: dict | None = None,
) -> None:
"""Mark dates around a controller boundary as missing local activity."""
current = _utc_datetime(current_ts)
current = as_utc_datetime(current_ts)
current_date = current.astimezone(ZoneInfo(current_tz)).date().isoformat()
_upsert_activity_day(conn, current_date, current_tz, incomplete=True)
if previous is None or not previous.get("local_tz"):
return
previous_at = _utc_datetime(previous["ts"])
previous_at = as_utc_datetime(previous["ts"])
previous_tz = previous["local_tz"]
previous_date = previous_at.astimezone(ZoneInfo(previous_tz)).date().isoformat()
_upsert_activity_day(conn, previous_date, previous_tz, incomplete=True)
@@ -1127,7 +1118,7 @@ def query_local_day_summary(
== recorded_date
)
day_seconds = int(
(_utc_datetime(utc_end) - _utc_datetime(utc_start)).total_seconds()
(as_utc_datetime(utc_end) - as_utc_datetime(utc_start)).total_seconds()
)
interval_seconds = activity_seconds or 0
full_activity_day = interval_seconds >= day_seconds or (