Too many alerts are as dangerous as too few: people stop reading them. Monitoring and alerting projects are about making every alert worth waking someone for.
Typical work
- Health and performance monitoring for infrastructure, platforms and applications
- Alert rules tied to user impact and service level objectives rather than raw thresholds
- On-call routing and escalation, integrated with your incident tooling
- Runbooks linked from alerts, so the first responder knows where to start
- Regular review of alert volume to remove noise
Our approach
We start with what your users would notice and work down from there. The result is a smaller set of alerts that people trust.
See also Observability and our On-Call Engineers (24 x7) service.