Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.ewake.ai/llms.txt

Use this file to discover all available pages before exploring further.

Ewake is built on one core idea: before you can resolve an incident, you need a complete picture of your production environment. Most teams don’t have that picture, every service exists, every signal is collected, but no human can hold all of it in their head. Ewake closes that gap. This page describes the concepts that drive how it works.

​
The Production Cartographer

Ewake’s central role is to draw the map of your production before it breaks, then keep it up to date as your environment evolves. Concretely, this means a continuous live scan of every node in your stack, services, dependencies, deployments, alerts, ownership, recent changes, assembled into a single coherent view. When an alert fires or a question is asked, ewake doesn’t start from zero: it already knows the topology of what’s affected and what could be related. This is the difference between “there’s an error spike on payments-api” and “this looks like the same failure pattern as incident #47, caused by a deploy on the dependency auth-service two hours ago.”

​
How ewake reasons

Ewake operates on a continuous loop:

Signal ingestion

Reads from your connected sources, Datadog, GitHub, Incident.io, Slack, in real time.

Context lookup

When a trigger occurs (alert fires, question asked), ewake cross-references the signal against the knowledge map.

Hypothesis ranking

Produces a ranked list of probable root causes, each with supporting evidence, recent commits, log excerpts, metric correlations, similar past incidents.

Output delivery

Returns the structured investigation to where you already work, Slack thread or Web App.

Learning loop

Each resolved incident, confirmed hypothesis, and engineer correction feeds back into the knowledge map.

​
Suggestion-first, not autonomous

Ewake is a force multiplier for the on-call human, not a replacement. Every output is a ranked hypothesis with a confidence score and supporting evidence. The engineer validates, decides, and acts. This is a deliberate design choice:
  • You see why, every hypothesis links to the evidence that produced it
  • You can override, if ewake’s top hypothesis is wrong, your correction is captured and improves the next investigation
  • You stay in control, ewake never silences alerts, restarts services, or modifies code without explicit human approval
The result: faster investigations without losing the human judgement that production reliability requires.