feat(alerts): the metric catalog is written by what emits it (#5061)

#5061 "Metric catalog belongs to what emits: plugins register their
metrics; names match what is recorded".

- The static METRIC_CATALOG is gone. metric_catalog() is core's own sources
  (the four monitor types and "steward", the app's open_fds/open_fds_pct,
  which were recorded but never offered) plus each loaded plugin's
  get_alert_metrics() export. A disabled plugin's sources leave the picker.
- traefik offers what it records: error_rate_4xx_pct and error_rate_5xx_pct.
  The old "error_rate" was never recorded, so a rule on it could not fire.
- tests/core/test_metric_catalog.py walks every MetricBatch.add /
  record_metric call (traefik's tuple loop included) and fails on a recorded
  metric the catalog lacks; it shows on a synthetic writer that it can fail.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
2026-10-06 12:55:55 -04:00
co-authored by Claude Opus 5.5
parent 58bfc7c361
commit 0742b7a2b7
10 changed files with 197 additions and 37 deletions
+8 -1
View File
@@ -26,7 +26,14 @@ ADMIN_TOOLS = {"delete_monitor", "delete_alert_rule"}
@pytest.fixture
def known(monkeypatch):
"""A source with one resource and its catalog metrics, without a database."""
"""A source with one resource and its catalog metrics, without a database:
the host_agent plugin is loaded (its get_alert_metrics hook writes the
catalog) and has reported one host."""
import plugins.host_agent
from steward.core import plugin_manager
monkeypatch.setattr(plugin_manager, "get_loaded_plugin_modules",
lambda: {"host_agent": plugins.host_agent})
async def resources(db, src):
return ["nas"] if src == "host_agent" else []
monkeypatch.setattr(alert_svc, "resources_for", resources)