Operations · Monitoring

Alert rules

Get paged before deliverability burns — thresholds on bounce rates, queues, circuits and vendor health, routed to email, Slack or PagerDuty.

Active rules
across 3 channels
Fired (7 d)
both auto-resolved
Silenced
maintenance window
MTTR
min
▼ 6 min vs last month

Rules

RuleConditionSeverityNotifyStatus
Bounce rate spike
any vendor, 15 min window
bounce_rate > 5% for 15m Critical PagerDuty · #mailops-ops
Complaint rate high
spam complaints, daily
complaint_rate > 0.1% / day Warning Slack #mailops-ops
Webhook queue lagging
event processing backlog
queue_depth > 1000 for 10m Critical PagerDuty · email on-call
Circuit open too long
vendor failover running
circuit_open > 30m Warning Slack #mailops-ops
Vendor API errors
5xx from any vendor
vendor_5xx > 2% for 5m Warning Slack #mailops-ops
Diagnosis backlog
LLM queue awaiting review
pending_diagnoses > 50 Info Email digest

Notification channels

S
Slack · #mailops-ops
Warnings + critical
Connected
PD
PagerDuty · MailOps on-call
Critical only
Connected
@
Email · on-call rota
Info + digest
Default

Recent alerts

Webhook queue lagging Critical
Depth hit 1,240 after a SendGrid event burst — autoscaling drained it.
Fired Aug 11 · 14:02 · auto-resolved in 22 min · paged on-call
Vendor API errors Warning
Scaleway TEM 5xx at 3.1% for 7 min — circuit opened, failover to SES.
Fired Jul 28 · 09:44 · auto-resolved in 41 min · posted to Slack
No open incidents
Everything below threshold for 16 days straight.

Open system health →