feat(settings): configurable monitoring thresholds

Move the hardcoded warn/crit cutoffs into Settings -> Thresholds (DB-backed,
live, no restart). New thresholds.* keys + to_thresholds_cfg() + a
threshold_style(value, kind) jinja global that reads them; latency reuses the
existing ping good/warn keys, uptime is direction-aware (floors).

Replace the _macros metric_style/uptime_style macros (now removed) with the
global across Hosts-Overview, host_agent fleet + panel, Uptime/SLA widget, and
the ping page uptime column — all now honor the configured cutoffs. Uptime keeps
its green 'good' look when not degraded.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

steward/templates/_macros.html

@@ -17,24 +17,6 @@
 </nav>
 {%- endmacro %}
 
-{# Threshold emphasis for a numeric metric (CPU/mem/disk %, etc).
-   Returns an inline-style fragment to drop into a span's style="": amber +
-   semibold at >= warn, red + bold at >= crit, empty (inherit) when the value
-   is normal or None. This makes the FAILING METRIC ITSELF stand out — not just
-   the status dot. Defaults suit percentage gauges (warn 80 / crit 90); pass
-   explicit warn/crit for other scales. Keep thresholds here so they live in one
-   place rather than hardcoded per template. #}
-{% macro metric_style(value, warn=80, crit=90) -%}
-{%- if value is not none and value >= crit -%}color:var(--red);font-weight:700;
-{%- elif value is not none and value >= warn -%}color:var(--yellow);font-weight:600;
-{%- endif -%}
-{%- endmacro %}
-
-{# Inverted emphasis for "higher is better" metrics like uptime %: amber below
-   `warn`, red below `crit`. The mirror of metric_style — same one-place-for-
-   thresholds discipline, opposite direction. Defaults suit SLA percentages. #}
-{% macro uptime_style(value, warn=99.0, crit=95.0) -%}
-{%- if value is not none and value < crit -%}color:var(--red);font-weight:700;
-{%- elif value is not none and value < warn -%}color:var(--yellow);font-weight:600;
-{%- endif -%}
-{%- endmacro %}
+{# Metric threshold coloring moved to the `threshold_style(value, kind)` jinja
+   global (steward/core/settings.py + app.py), which reads operator-configurable
+   cutoffs from Settings → Thresholds. Call it as threshold_style(v, 'cpu'). #}