feat(hosts): one health header, graded by one set of thresholds (#5122)

The host page's vitals strip becomes its health header: Live or Stale,
then one sentence — "Memory 93% is past its critical threshold", "Every
reading is within its thresholds", or, for a stale host, "No report since
20m ago; the readings below are from then" — and last seen as |ago. CPU,
memory, root disk and load per core follow as stat tiles graded by
Settings → Thresholds, with their sparklines.

The full-metrics page drew its bars and gauges at a hard-coded 70/90
(temperatures 70/85) while every other page used the configured cutoffs
(80/90 by default), so a 75 % CPU was amber here and fine everywhere else.
bar() now takes the metric kind and colours its .meter-fill through
status_class; gauges, temperatures and per-core bars do the same.

The host's monitor rows move onto the Monitors list's shapes: a status dot,
the name linking to the monitor's page, Ping/TCP/DNS/HTTP, pill-up/
pill-down, latency and uptime through status_class, labelled icon buttons.
HEX_CEILING drops to 54.

New formatting.worst_reading(readings, thresholds) picks the reading
furthest past its cutoff, the first of equals winning.

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

tests/plugins/host_agent/test_host_health.py

@@ -0,0 +1,79 @@
+"""The host page says how the host is in one header, and grades every reading by
+Settings → Thresholds, nowhere else (#5122)."""
+import re
+from datetime import datetime, timedelta, timezone
+from types import SimpleNamespace
+
+import pytest
+from quart import render_template
+
+from steward.core.formatting import worst_reading
+from steward.core.settings import to_thresholds_cfg
+from tests.test_style_tokens import hex_colours
+
+TH = to_thresholds_cfg({})  # cpu/mem/disk warn 80 crit 90; load warn 80 crit 100
+
+
+@pytest.mark.parametrize("readings, worst", [
+    ([("CPU", 40, "cpu"), ("Memory", 50, "mem")], None),
+    ([("CPU", None, "cpu")], None),                                       # no reading is not a problem
+    ([("CPU", 85, "cpu"), ("Memory", 95, "mem")], ("Memory", 95, "crit")),  # crit beats warn
+    ([("CPU", 92, "cpu"), ("Memory", 95, "mem")], ("CPU", 92, "crit")),     # first of equals
+    ([("Disk /", 81, "disk")], ("Disk /", 81, "warn")),
+])
+def test_worst_reading(readings, worst):
+    assert worst_reading(readings, TH) == worst
+
+
+async def _vitals(app, *, cpu=40.0, mem=50.0, stale=False, seen_min=1):
+    readings = (("CPU", cpu, "cpu"), ("Memory", mem, "mem"), ("Disk /", 30.0, "disk"),
+                ("Load per core", 20, "load"))
+    async with app.test_request_context("/plugins/host_agent/vitals/h1"):
+        page = await render_template(
+            "_host_vitals.html", reporting=True, stale=stale,
+            reg=SimpleNamespace(last_seen_at=datetime.now(timezone.utc) - timedelta(minutes=seen_min),
+                                agent_version="1.9.0"),
+            cpu=cpu, mem=mem, disk_root=30.0, load_per_core=20,
+            sparks={"cpu": "", "mem": "", "disk": "", "load": ""},
+            psi={"cpu": None, "mem": None, "io": None}, worst=worst_reading(readings, TH))
+    return re.sub(r">\s+<", "><", " ".join(page.split()))
+
+
+async def test_a_host_over_a_threshold_says_which_reading(app):
+    page = await _vitals(app, cpu=40.0, mem=93.0)
+    assert '<span class="badge badge-ok">Live</span>' in page
+    assert '<span class="tone-crit">Memory 93%</span> is past its critical threshold.' in page
+    assert 'class="stat-tile stat-crit"><span class="stat-tile-label">Memory</span>' in page
+    assert 'class="stat-tile stat-ok"><span class="stat-tile-label">CPU</span>' in page
+    assert "seen <time" in page and "agent v1.9.0" in page
+    assert not hex_colours(page), hex_colours(page)
+
+
+async def test_a_quiet_host_says_so(app):
+    assert "Every reading is within its thresholds." in await _vitals(app)
+
+
+async def test_a_stale_host_says_since_when_before_any_reading(app):
+    page = await _vitals(app, mem=93.0, stale=True, seen_min=20)
+    assert '<span class="badge badge-warn">Stale</span>' in page
+    assert "No report since <time" in page and ">20m ago</time>; the readings below are from then." in page
+    assert "past its" not in page          # an old reading is not news
+
+
+async def test_the_metrics_page_grades_by_the_same_cutoffs(app):
+    # 75 % was amber under the old hard-coded 70/90; Settings → Thresholds says warn at 80.
+    async with app.test_request_context("/plugins/host_agent/h1/metrics"):
+        page = await render_template(
+            "_host_metrics.html", host=SimpleNamespace(address="10.0.0.2"), stale=False,
+            reg=SimpleNamespace(distro=None, kernel=None, arch=None, agent_version=None,
+                                last_seen_at=datetime.now(timezone.utc)),
+            hostlvl={"cpu_pct": 75.0, "mem_used_pct": 95.0, "temp_c_max": 72.0},
+            cores=[(0, 85.0)], nets={}, disks_io={}, temps={"cpu": 90.0},
+            mounts={"/": {"disk_used_pct": 50.0, "disk_used_bytes": 1, "disk_total_bytes": 2}})
+    assert 'class="stat-val tone-ok" style="font-size:1.6rem;">75%' in page
+    assert 'class="stat-val tone-crit" style="font-size:1.6rem;">95%' in page
+    assert 'class="meter-fill meter-ok" style="width:75.0%;"' in page
+    assert 'class="meter-fill meter-warn" style="width:85.0%;"' in page   # the core
+    assert 'class="stat-val tone-warn" style="font-size:1.6rem;">72°C' in page
+    assert '<span class="tone-crit" style="font-variant-numeric:tabular-nums;">90°C' in page
+    assert not hex_colours(page), hex_colours(page)