Skip to main content
This page is the AGENTS.md for ewake, a context file written for AI coding assistants and external agents (Cursor, Claude Code, Codex) that connect to ewake. If you are an AI agent, read this page in full before issuing tool calls. It tells you what ewake is, what you can ask, what you cannot, and how to query it well.

​
What ewake is

Ewake maintains a live knowledge map of a customer’s production environment by ingesting data from observability tools (Datadog, Grafana, Prometheus), source control (GitHub, GitLab), incident management (Incident.io, Rootly, FireHydrant), and Slack. Ewake is an on-call agent that takes the page first. It investigates alerts, deploys and scheduled tasks, produces ranked root-cause hypotheses with supporting evidence, and takes action on the customer’s behalf, such as restarting a pod or merging a PR, while keeping their team in the loop. Only issues that need a human decision are escalated. For the conceptual model, see Concepts.

​
How to interact with ewake

There are two stable interaction surfaces for agents: You should never scrape the Web App or Slack interface. Use one of the two surfaces above.

​
Authentication

All access requires an ewake API key, scoped to a single customer workspace.
API keys are generated in the ewake dashboard under API Keys in the sidebar. A key is scoped to one customer workspace and cannot reach another customer’s data. Within that workspace it is not scoped further: the same key authenticates deployment events and the MCP server. Treat it as a full workspace credential.

​
Capabilities, what you can ask

Ewake answers questions about a customer’s production environment. Common patterns:
  • Service health, current state, error rates, latency for any named service
  • Recent activity, deployments, alerts, incidents within a time window
  • Correlation, link an alert or anomaly to recent changes (deploys, commits, infrastructure)
  • Historical context, has this issue happened before? What resolved it?
  • Dependency reasoning, upstream/downstream impact of a change
  • Investigation, given an alert or symptom, produce ranked hypotheses with evidence
For ready-made example queries, see What can I ask ewake?.

​
Limitations, what you cannot do

These are hard limits on what you can obtain through this API. Do not attempt them.
  • No actions through this API, ewake takes actions (such as restarting a pod or merging a PR) as part of its own on-call work for the customer’s team. You cannot trigger, request, or change an action through the MCP server or REST API; both are read-only for you.
  • No cross-customer queries, your API key is workspace-scoped. Asking about another customer’s data will not work.
  • No model training on customer data, ewake does not feed customer signals into foundation-model training.
  • No raw telemetry retention, logs and metrics are queried live; ewake cannot return data older than what the underlying source retains.
  • No escalation through this API, ewake escalates issues that need a human decision to the customer’s team, through the channels they configured. You cannot page or notify anyone through this API.
For the full security and access matrix, see Permissions & Data Access.

​
Conventions

Follow these conventions when querying ewake, they materially improve answer quality.
  • Use the exact service name as it appears in Datadog (or your primary observability tool). ewake correlates by service identifier; “checkout API” and “checkout-api” are not the same.
  • Specify a time window for any historical question. Default lookbacks vary; explicit windows (“past 2 hours”, “since the deploy at 14:30 UTC”) produce more precise answers.
  • Ask one question per call. Compound questions (“why is X slow AND who deployed Y”) fragment ewake’s reasoning. Issue them as separate tool calls.
  • Treat output as hypotheses, not facts. Every ewake answer includes a confidence indicator and supporting evidence. Surface both to the human user.

​
Domain vocabulary

When parsing ewake responses, these terms have specific meanings:

​
Example interaction patterns

Good, focused, specific, time-bounded:
Avoid, vague, multi-question, no time window:
Good, asking for evidence-backed reasoning:

​
Versioning

This page reflects ewake’s current capabilities. Capabilities expand over time. If you cache this content, refresh at least every 30 days. For questions or to report incorrect behaviour, contact support@ewake.ai.