Sentinel

On-call incident copilot

Forty alerts, one incident, and what fixed it last time.

Sentinel takes your alert emails, collapses the storm into one incident card, and searches your own resolved incidents for one that looks similar — then shows you the action item a person wrote when it happened before.

It suggests from history. It does not diagnose. Every suggestion cites a prior incident by date, and when nothing in your library is close enough the screen says so rather than guessing.

Three steps, and the third one is the product

1 · Forward the alerts

Two doors. Each workspace gets a webhook URL to POST to — Alertmanager and Grafana work as they are — and an address at in.sentinel.deplyra.com for tools that only send email.

2 · One card, not forty

Repeats of the same check collapse for two hours; different symptoms naming the same service collapse for twenty minutes. The rule that fired is printed on every alert, so you can disagree with it.

3 · Similar to 3 Mar

When an incident resembles one you have resolved, you get the earlier incident, its date, the words the two share, and the action item you wrote. If nothing is close, you get told nothing is close.

What it is not

Sentinel does not detect anything and will not wake anyone up. It sits downstream of the tool that pages you. It holds no shared corpus of other companies' outages: the library it searches is your own history and nothing else, so it is useless on your first day and useful on your fiftieth.

It will not tell you what broke. A tool that offered a cause from the text of an alert would be guessing, and a confident guess at 3am is worse than silence — you would chase it.

Pricing

Nothing is charged. There is no card form on this site and no paid plan behind it. If you want to know when there is one, leave an address.

No card, no charge, no trial clock.