We go on call in week one, then work every quarter to make your alerts rarer.

A 20-minute call and a one-page service charter

We look at your alerts and on-call together and tell you if we are a fit and what it costs. If we are, we agree which services we cover, what is out of scope, who on your side makes the call during an incident and how we measure success. Both are free.

We go on call. Cover and billing start.

Cover and billing start

You create one read-only identity from our template and point your alerts at us. From then on, every alert is acknowledged within 15 minutes and assessed within 30.

You do: create the identity, connect your alerts, add us to a Slack channel and ticket queue.

We learn your systems

Your team walks us through the architecture. We read your five most recent incidents, trace how requests move between services and write down who owns each layer: your code, your cloud provider and any third-party software.

You do: one architecture walkthrough, and choose whether to share code repositories.

You get a written gap report

Each way production could fail without anyone being told, or fail in a way nobody could quickly undo: user journeys no alert would catch, alerts that wake people for nothing, databases without a tested restore, services in only one zone, releases that cannot be rolled back. Each gap comes with what it would cost you and how much work it is to close.

You do: choose which gaps we close first.

A starter pack of runbooks for your approval

Runbooks are step-by-step fixes for known issues: rolling back a release, adding capacity, restarting a service, moving traffic away from a failing zone. Each is tested before it can run in a real incident.

You do: approve the runbooks you want, and apply each one’s narrow permission to the identity yourself.

Reliability targets set, and alerts rebuilt around them

Your business owners decide how reliable your most important service needs to be, such as 99.9% of checkouts succeeding in a month. We rebuild its alerts around that target. Where a gap needs a change to your code or infrastructure, we write it as a change your team reviews and ships.

Heads up: better alerting can make the first month look busier, because problems that were always there start showing up. We tell you which ones.

Report, review, next gaps

Every month

Each month we walk your team through the report: response times against the contract, slowest included, how often our first likely cause was right, every incident and the follow-ups still open. Each quarter we agree one or two reliability goals, each with a target and a date. The number of alerts that need a person should fall over time.

See the sample monthly report

No lock-in after month three

Cover is billed monthly. After the first three months, either side can end it with 30 days’ notice. The runbooks and documentation we wrote stay with you.