max / makenotwork

fix: suppress spurious 'recovered' alert on monitor startup The health monitor initializes previous_status as None, so the first successful check was treated as a None->Operational transition and fired a 'MNW recovered — all services operational' email on every clean restart. Skip the alert (and WAM ticket / subscriber notifications) for the bootstrap case while still recording previous_status so the next real degradation alerts normally.

Author: Max J. <87768334+MaxJMath@users.noreply.github.com> · 2026-05-21 03:19 UTC

Commit: 0467b78bf2b8a7ce50674542ba63564d1669826f

Parent: 39d5998

1 file changed, +7 insertions, -4 deletions

M server/src/monitor.rs +7 -4

			@@ -131,9 +131,12 @@ pub fn spawn_monitor(
131	131		// Update Prometheus DB pool gauges on every tick
132	132		crate::metrics::record_db_pool_stats(&state.db);
133	133
134		-	// Log status changes
	134	+	// Log status changes. Skip the bootstrap None->Operational transition
	135	+	// so clean restarts don't fire a spurious "recovered" alert.
135	136		let status_changed = previous_status != Some(snap.status);
136		-	if status_changed {
	137	+	let is_bootstrap_ok =
	138	+	previous_status.is_none() && snap.status == MonitorStatus::Operational;
	139	+	if status_changed && !is_bootstrap_ok {
137	140		match snap.status {
138	141		MonitorStatus::Operational => {
139	142		if previous_status.is_some() {
			@@ -222,10 +225,10 @@ pub fn spawn_monitor(
222	225		);
223	226		wam.create_ticket(&title, Some(&body), priority, "health-status-change", None).await;
224	227		}
225		-
226		-	previous_status = Some(snap.status);
227	228		}
228	229
	230	+	previous_status = Some(snap.status);
	231	+
229	232		// DB pool pressure check (>80% active connections), with cooldown
230	233		{
231	234		let pool_size = state.db.size();