Skip to content
DashboardWebsite

Notifications

Just want to do it in the dashboard? See the guide: Get alerts in Slack, by email or on a webhook.

A pull request’s report appears as a PR comment and a commit status. Notification channels send analysis events to other destinations, such as a Slack or Teams channel, organization members’ inboxes, or an endpoint you manage. This page explains how to connect channels and what they receive.

An organization can have up to 20 channels by default (see Operator settings). Each channel can subscribe to selected events and, optionally, serve only some teams (see Team channels).

KindDestinationSigning
slackAn Incoming Webhook URL from an app in your workspaceSlack uses the URL to authenticate
teamsA Workflows webhook URL for a Teams channelThe URL authenticates the request
emailSelected members of the organization, through the deployment’s mail serverNot applicable
webhookAn HTTPS endpoint that receives the event payload as JSONOptional HMAC-SHA256

In the dashboard, open Organization → Connections → Notifications. Each tab lists channels of one kind. The add button and each row’s Edit button open the channel form in a side panel.

Discord and PagerDuty are not supported.

You need a Slack app in your workspace. An incoming webhook belongs to an app. OpenTremor does not register an app with Slack or access your workspace.

Use the guided flow if this deployment has a public URL. The manual flow also works for air-gapped deployments.

Set up the app once. Then install one channel at a time through Slack.

  1. In the dashboard, open Organization → Connections → Notifications, select the Slack tab, and click Show the app manifest. Copy the JSON.

  2. At api.slack.com/apps, choose Create New App → From a manifest, select your workspace, and paste the JSON. The manifest enables incoming webhooks and registers this deployment’s redirect URL.

  3. On the app’s Basic Information page, copy the Client ID and Client Secret into the dashboard, then save. The secret is encrypted at rest and is not shown again.

  4. Click Add to Slack. Choose a channel on Slack’s screen and click Allow. You return to the Slack tab, where OpenTremor creates a channel named after the Slack channel.

  5. Select the events you want. The new channel has no event subscriptions until you select them.

  6. Click Test to check delivery.

Repeat step 4 for each additional Slack channel. Slack allows one channel per install.

The guided flow requires server.public_base_url. Slack redirects a browser to this server, which exchanges a code with slack.com. If the deployment cannot reach Slack, use the manual flow.

  1. Create a Slack app at api.slack.com/apps → Create New App → From scratch. Choose the workspace where you want to post.

  2. Enable incoming webhooks. In the app settings, open Incoming Webhooks and turn on Activate Incoming Webhooks.

  3. Add a webhook to the workspace. Click Add New Webhook to Workspace and choose a channel. Slack returns a URL like https://hooks.slack.com/services/T000/B000/xxxxxxxx. The URL is bound to that channel. Repeat this step for each additional channel.

  4. In the dashboard, open Organization → Connections → Notifications, select the Slack tab, and paste the URL under Add a Slack channel by URL. Give the channel a label so you can distinguish it later; the URL is not displayed again.

  5. Select the events you want, then click Test.

Teams receives notifications through a Workflows webhook, with one URL per channel.

  1. In Teams, open the channel, select More options (…) → Workflows, and choose Post to a channel when a webhook request is received.
  2. Name the workflow, confirm the team and channel, then select Add workflow. Copy the URL Teams shows.
  3. In the dashboard, open Organization → Connections → Notifications, select the Microsoft Teams tab, and paste the URL. Give the channel a label.
  4. Select the events you want, then click Test. Teams returns 202 Accepted before the workflow runs, so check the channel to confirm that the card arrived.

Each event arrives as an Adaptive Card posted through the Workflows app on behalf of the workflow’s owner. The card lists up to 5 items, then links to the dashboard.

A Teams workflow URL is a credential: anyone holding it can post to its channel. OpenTremor encrypts the URL and never returns it. If it leaks, turn off or delete the workflow in Teams.

An email channel sends each event to selected organization members. It has no URL.

  1. In the dashboard, open Organization → Connections → Notifications, then select the Email tab.
  2. Under Add an email channel, give the channel a label and select the members who should receive messages.
  3. Select the events you want, then click Test. Each recipient gets one test message.

An email channel stores member IDs, not email addresses. Only active organization members can receive mail. A member who leaves, or whose membership or account is deactivated, stops receiving mail. A changed email address is used automatically. To reach a mailing list, point a webhook channel at your own mail relay.

A channel can have at most 20 recipients by default (notifications.max_recipients_per_channel). Each recipient gets one plain-text message per event with the title, summary, up to 20 items, and links.

A channel can subscribe only to events the deployment can emit. GET /v1/notifications/events returns the event catalogue that the dashboard uses to build its event switches.

EventFires when
pr.gate_blockedAn analysis sets a pull request’s check to failure or pending.
finding.needs_reviewA finding enters needs_review because of low model confidence, a rule marked requires_review, or a manual triage action.

New channels created through the manual channel form subscribe to both events by default. Slack channels created through the OAuth install flow start with no subscriptions; select events after installation.

pr.gate_blocked fires only when the check does not pass. There is no event for every successful analysis.

A channel is org-wide by default and receives every event it subscribes to. Set team_ids (in the dashboard, Only for some teams) to receive events only when they concern one of those teams:

EventTeams associated with the event
pr.gate_blockedTeams that review the pull request’s repository
finding.needs_reviewTeams assigned to the finding

Repository review teams are set on the Repositories page; see Review routing.

  • An event with no team goes only to org-wide channels.
  • An org-wide channel receives every event, including events also sent to a team channel.
  • When a team is deleted, it is removed from the channel. If all of a channel’s teams are deleted, the channel receives nothing; it does not become org-wide.

A Slack channel can also mention a user group (slack_mention), using an ID like S0123ABC from the group’s page in Slack. The mention appears above the message and notifies the group’s members.

Every channel kind renders from the same envelope. A webhook channel receives it as JSON:

{
"event": "pr.gate_blocked",
"org_id": "org_...",
"timestamp": "2026-09-09T12:34:56.000000+00:00",
"title": "acme/billing#412: check failure — 2 blocking findings",
"severity": "CRITICAL",
"summary": "1 CRITICAL, 1 HIGH across 12 analysed resources",
"subject": "acme/billing#412",
"links": [
{"label": "Triage in the dashboard", "url": "https://dashboard.example.com/findings?changeset=acme%2Fbilling%23412"},
{"label": "Open the pull request", "url": "https://github.com/acme/billing/pull/412"}
],
"items": [
{
"id": "PYCC-001",
"severity": "CRITICAL",
"title": "Invoice export scoped by an org_id from the query string",
"location": "billing/api/invoices.py:export_invoices"
}
],
"counts": {"critical": 1, "high": 1, "medium": 0, "low": 0, "resources": 12},
"team_ids": ["payments"]
}
  • items contains findings. id is the rule ID and location is the resource address. The payload includes up to 10 items. Slack and Teams list up to 5 before referring readers to the dashboard. counts covers the full set.
  • team_ids lists the teams associated with the event. An empty list means the event concerns the organization as a whole. Team channels match against this list.
  • Findings are post-triage. Suppressed findings and false positives are removed, so a notification does not report a finding a reviewer dismissed.
  • A link is omitted when its base URL is not configured; it is not sent with a null URL. See server.public_base_url and server.dashboard_base_url in the reference.
  • title, summary, and strings under items can contain model output derived from the content under analysis. Escape them when rendering. OpenTremor escapes text for its own destinations; a raw webhook payload receives JSON encoding only.

When a webhook channel has a configured secret, each request includes X-Signature-256: sha256=<hex>. The value is HMAC-SHA256 over the exact request body, using the same scheme GitHub uses for inbound webhooks:

import hashlib, hmac
def verify(raw_body: bytes, header: str, secret: str) -> bool:
expected = "sha256=" + hmac.new(secret.encode(), raw_body, hashlib.sha256).hexdigest()
return hmac.compare_digest(expected, header)

Compute the signature over the raw bytes. Do not re-serialize parsed JSON before verifying; key order and whitespace can change.

Delivery is best-effort. A slow or failing destination does not fail or delay the analysis that triggered the event. There is no retry queue. Failed deliveries are recorded on the channel and discarded. Check the dashboard’s last delivery outcome to spot failures.

If you need to retain every notification, subscribe a webhook channel pointed at a system you control, then put each event in a durable queue there.

Set these values in a configuration file or environment variable. You cannot edit them in the product. The egress policy controls which destinations the server can send findings to. Link-local addresses, where cloud metadata lives, remain blocked.

notifications:
egress:
enabled: true # false disables notification channels entirely
allow_private_networks: false # true to reach a destination inside your own network
allowed_hosts: [] # exact hostnames or "*.suffix"; empty means any permitted address
max_channels_per_org: 20
max_recipients_per_channel: 20
timeout_seconds: 5.0

Set server.public_base_url. Slack redirects the browser to <public_base_url>/notifications/slack/callback, and the manifest registers that URL on your app. This server also needs outbound access to slack.com to exchange the authorization code.

These calls use fixed hosts, not customer-supplied addresses, so notifications.egress does not apply. That policy governs destinations an organization configures. A deployment that blocks outbound access to slack.com cannot use the guided flow; use the manual setup.

Each organization provides its own Slack app. Operators do not need to register an app or obtain a Slack Marketplace listing or review.

notifications.egress is separate from llm.egress because the two policies control different destinations. For example, a cloud instance can set llm.egress.enabled: false so organizations use hosted provider endpoints, while still allowing delivery to hooks.slack.com. A shared policy would tie permission to use a custom model to permission to send findings to an internal address.

By default, the policy allows public addresses and blocks private networks. A self-hosted deployment that sends notifications to an internal Mattermost or webhook must set allow_private_networks: true. Link-local addresses (169.254.0.0/16, fe80::/10) remain blocked because cloud instance metadata uses them.

The system checks each destination URL when it is saved and before each delivery. The save-time check gives an explanatory error for an invalid destination. The delivery-time check catches a hostname that resolves to a different address after saving.