CloudWatch Alarm Routing to On-Call

CloudWatch alarms hit an SNS topic subscribed to LunoGen. The agent checks the current on-call rotation, adds runbook links, and DMs the right engineer on WhatsApp — not just the #alerts channel.

Webhook + WhatsApp DMcloudwatchon-callalarms

Conversation example

SNS message received: ALARM rds-prod-high-cpu (value 94% > 85%)
memory_recall

CloudWatch alarm firing on RDS prod. Looking up the current on-call rotation and recent deploys...

whatsapp_web_send

CloudWatch Alarm — rds-prod-high-cpu Resource: lunogen-prod-db (db.m6g.2xlarge) Metric: CPUUtilization = 94% (threshold 85% for 3 datapoints) Region: us-east-1 Severity: P2 (degraded, not down) Duration: 14 minutes Context: - No deploy in the last 6 hours - Connection count normal - Slow query log: 'SELECT * FROM events WHERE ...' (no index) On-call: @sam (primary) — DM'd on WhatsApp with runbook link: https://docs.lunogen.com/runbook/rds-cpu Backup: @daniel-p (notified in #oncall-backup) I'll escalate if unacked in 5 min.

More AWS use cases

Deploy this in minutes

Create a LunoGen agent, connect AWS, and start running this workflow from WhatsApp today.

Back to AWS