perf(host_agent): the sensor judgment reads the week once, then only new hours (#5786)

On the live instance the hourly refresh took 3 min 8 s, averaging a week of
raw temp_c samples each time. The hourly averages are now kept in memory.
The first refresh after startup reads the whole window. Each later one reads
raw samples from the start of the hour the last read ended in, so the hour
that was still filling is read again whole. merge_hours folds the new rows
in and drops hours older than the window.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

tests/plugins/host_agent/test_sensors.py

@@ -72,3 +72,13 @@ def test_a_judged_sensor_is_left_out_of_the_host_temperature_and_listed_with_why
     real, unwired = split_temps("util", {"temp3": 62.0, "Package id 0": 55.0})
     assert list(real) == ["Package id 0"]
     assert unwired == {"temp3": (62.0, "falls as the CPU warms (r -0.81 over 168 h)")}
+
+
+def test_a_later_read_replaces_the_hour_that_was_filling_and_old_hours_drop():
+    hours = {"h:temp:a": {T0: 50.0, T0 + timedelta(hours=1): 51.0}}
+    s.merge_hours(hours, [("h:temp:a", T0 + timedelta(hours=1), 55.0),     # the hour, read whole
+                          ("h:temp:a", T0 + timedelta(hours=2), 56.0),
+                          ("h:temp:b", T0 + timedelta(hours=2), None)], since=T0 + timedelta(hours=1))
+    assert hours == {"h:temp:a": {T0 + timedelta(hours=1): 55.0, T0 + timedelta(hours=2): 56.0}}
+    s.merge_hours(hours, [], since=T0 + timedelta(hours=9))                # a sensor gone quiet
+    assert hours == {}