fix(alerts): a metric picked on one part doesn't follow you to another (#5475)

#5475 "Rule form: pick the host, then its part…". CI run 8909 caught it:
moving from the whole host to a mount kept "CPU" in the mount's Metric list,
because a chosen metric always stayed on offer. Only a rule's own page keeps
its resource and metric after they go quiet (keep=True from edit_rule); the
pickers' swaps drop a metric the new part doesn't report.

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

steward/alerts/routes.py

@@ -295,10 +295,12 @@ async def cover_groups_api():
 
 
 async def _target(db, source: str, *, host: str = "", resource: str = "", metric: str = "",
-                  covers: str = "resource") -> dict:
-    """The rule form's target pickers and the chosen metric's preview (#5475)."""
+                  covers: str = "resource", keep: bool = False) -> dict:
+    """The rule form's target pickers and the chosen metric's preview (#5475);
+    `keep` on a rule's own page, so its resource and metric stay on offer."""
     from steward.core.status import sparkline_svg
-    target = await alert_svc.rule_target(db, source, host=host, resource=resource, metric=metric)
+    target = await alert_svc.rule_target(db, source, host=host, resource=resource, metric=metric,
+                                         keep=keep)
     preview = None
     if target["resource"] and target["metric"]:
         preview = await alert_svc.metric_preview(db, source, target["resource"], target["metric"])
@@ -360,7 +362,7 @@ async def edit_rule(rule_id: str):
             return redirect(f"/alerts/rules/{rule.parent_id}/edit")
         ctx = {**await _cover_choices(db, rule.source_module, rule),
                **await _target(db, rule.source_module, resource=rule.resource_name,
-                               metric=rule.metric_name, covers=rule.covers)}
+                               metric=rule.metric_name, covers=rule.covers, keep=True)}
         hosts = sorted((c.resource_name for c in (await db.execute(
             select(AlertRule).where(AlertRule.parent_id == rule.id))).scalars()), key=str.lower)
         history_count = (await db.execute(

steward/alerts/service.py

@@ -391,13 +391,14 @@ def alertable(source_module: str, reported: set[str]) -> list[str]:
 
 
 async def rule_target(db, source_module: str, *, host: str = "", resource: str = "",
-                      metric: str = "") -> dict:
+                      metric: str = "", keep: bool = False) -> dict:
     """What the rule form's target pickers offer (#5475), from the latest-value
     table: the resources that report something a rule can watch, and for the
     chosen one the metrics it reports. For a host source the resources split
     into a Host picker and a Part picker ("the whole host", then the host's
-    mounts, cores and so on). A resource or metric already chosen stays on
-    offer even when it has gone quiet, so editing a rule never drops them.
+    mounts, cores and so on). With `keep` (a rule's own page), its resource
+    and metric stay on offer even after they go quiet, so opening a rule
+    never drops them; a metric carried over from another part is dropped.
 
     Returns host_source, hosts, host, parts [(heading, [(value, label)])],
     resources, resource, metrics, metric."""
@@ -434,7 +435,7 @@ async def rule_target(db, source_module: str, *, host: str = "", resource: str =
     else:
         out["resources"] = sorted(pickable, key=str.lower)
     metrics = list(pickable.get(out["resource"]) or catalog)
-    if metric and metric not in metrics and (metric in catalog or not catalog):
+    if keep and metric and metric not in metrics and (metric in catalog or not catalog):
         metrics.append(metric)
     out.update(metrics=metrics, metric=metric if metric in metrics else "")
     return out

tests/core/test_alert_rule_target.py

@@ -59,7 +59,14 @@ async def test_picking_another_host_starts_at_the_whole_host():
 
 async def test_a_rule_s_own_resource_stays_on_offer_after_it_goes_quiet():
     got = await alert_svc.rule_target(_db(), "host_agent", resource="h3:/data",
-                                      metric="disk_used_pct")
+                                      metric="disk_used_pct", keep=True)
     assert "h3" in got["hosts"] and got["host"] == "h3"
     assert got["parts"] == [("Mounts", [("h3:/data", "/data")])]
     assert got["metric"] == "disk_used_pct"
+    # A reporting part whose metric went quiet keeps it on the rule's own page.
+    got = await alert_svc.rule_target(_db(), "host_agent", resource="h1:/var", metric="disk_free_pct",
+                                      keep=True)
+    assert got["metric"] == ""                                     # not in the catalog: not offered
+    got = await alert_svc.rule_target(_db(), "host_agent", resource="h1:/var", metric="uptime_secs",
+                                      keep=True)
+    assert got["metrics"] == ["disk_used_pct", "uptime_secs"] and got["metric"] == "uptime_secs"