LandfallDocs
Integrations · Telemetry

Connecting Sentry to Landfall

Sentry is usually the first system to know an incident exists — most real incidents start life as an application exception, not an infrastructure alarm. This guide covers three things: creating a scoped auth token and pasting it in, what the investigator agent can read once connected, and — under its own section below — wiring Sentry's issue-alert webhook so it opens a Landfall war room automatically.

Sentry
Auth token
Landfall
Settings → Integrations

What you'll need

  • A Sentry account, and to know your organization's slug (shown in the URL after sign-in, e.g. sentry.io/organizations/<org-slug>/).
  • An admin on your Landfall organization, to save the connection under Settings → Integrations.
  • If you self-host Sentry: your instance's base URL — Landfall calls the same /api/0/… surface on any origin.

1. Create a Sentry auth token

Landfall only ever reads from Sentry: it verifies the token is alive and then reads issues/events when an incident is being investigated. An Internal Integration's own token is the narrowest option — scoped to one internal integration you create and name, rather than a broad personal auth token. Give it Read access to Issue & Event and Read access to Organization; nothing else is needed.

  1. Sentry → Settings → Developer Settings → New Internal Integration

    Name it something you'll recognize later (e.g. landfall-integration).

    sentry.io/settings/<org-slug>/developer-settings/new-internal/
    🖼Sentry's New Internal Integration form, named and with Issue & Event / Organization permissions set to Read.
    Screenshot pending — see this guide's tracked follow-up.
  2. Permissions: Issue & Event → Read, Organization → Read

    Both are read-level, nothing more. Landfall never resolves, assigns, or otherwise mutates a Sentry issue through this token — the read plugin's operations are structurally read-only regardless of what a broader token could do.

  3. Save, then copy the token

    Sentry shows the token once, on the integration's detail page after saving. Copy it now: you'll paste it straight into Landfall in the next step.

    sentry.io/settings/<org-slug>/developer-settings/landfall-integration/
    🖼Sentry showing a newly created Internal Integration's token value, with a copy-to-clipboard control.
    Screenshot pending — see this guide's tracked follow-up.

2. Add it to Landfall

In Landfall, open Settings → Integrations → Sentry for your organization. It asks for the token, your organization slug, and two optional fields: a default project slug (used when a read doesn't name one explicitly) and a self-hosted API base URL override.

  1. Paste the token and organization slug, then Test connection

    Landfall calls Sentry's own organization-details API with the token before saving anything — a token that can't authenticate, or an organization slug Sentry doesn't recognize, is refused outright rather than stored as "configured but unverified."

    Settings → Integrations → Sentry
    🖼Landfall's Sentry integration panel showing a green connected status and the message Sentry organization verified.
    Screenshot pending — see this guide's tracked follow-up.
  2. Save integration

    Sentry now shows Configured alongside your organization's other connectors.

3. Triggering Landfall from Sentry

Connecting Sentry (above) lets the investigator agent read your issues and events during an investigation. It's a separate step to have a Sentry issue-alert open that investigation in the first place — see Triggering Integrations → Sentry for the webhook setup, the exact body Sentry sends, and worked examples. The two are independent; most teams eventually set up both.

Test connection failed?
  • Double check the organization slug matches exactly what's in your Sentry URL; a typo is the most common cause.
  • Confirm the Internal Integration hasn't been disabled or the token regenerated since you copied it.
  • The token needs Read access to Organization at minimum — a token scoped to only Issue & Event will be rejected by the verify call.
  • Self-hosted Sentry: confirm the API base URL override has no trailing path beyond the origin (Landfall appends /api/0/… itself).
Integration connects, but the agent can't find any issues?
  • The connection test only checks organization-level read access. If the token is missing Issue & Event: Read, issue/event queries fail with a permissions error even though the integration shows Configured.
  • Edit the Internal Integration in Sentry and add the missing permission — no need to create a new token or reconnect in Landfall, the existing token keeps working.
  • Confirm a default project slug is set (or that the agent is being asked about a specific one) — listIssues needs a project.

Reference

FieldValue
Auth tokenA Sentry Internal Integration or Organization Auth Token, sent as Authorization: Bearer
Required scopesorg:read (connection check) + event:read (issue/event reads)
Organization slugYour Sentry org's slug — shown in the URL after sign-in
Default project slugOptional; used when a read doesn't name a project explicitly
API base overrideOptional; leave blank for Sentry Cloud (sentry.io), set only for self-hosted Sentry
Read operationsProject issues (listIssues) and an issue's error events (listIssueEvents) — stack traces, tags, breadcrumbs, not performance/tracing data. full=true on listIssueEvents returns the complete event payload, which can include request/user context; it is returned exactly as Sentry sends it, never filtered.