Notification Rules

Notification rules decide which Stacktape events an alert channel receives, such as a failed deployment, an opened incident or a changed secret. Each rule belongs to one channel, lists one or more event types and can be limited to specific projects and stages. Rules are managed on the channel's page in the Console.

Notifications cover events recorded by Stacktape: deployments, deletions, rollbacks, Git deployments, script runs, secret and SSM parameter changes, incidents and new security findings. For metric-based alerts on running resources (CPU, error rates, latency), use alarms. A firing alarm opens an incident, and a notification rule that watches incident events sends it to your channel.

When to use notifications

Notification rules are most useful when your team needs to know about deployment activity or incidents without watching the CLI or Console. Common scenarios:

  • Production deploy alerts. Notify a Slack channel when a production deploy fails, so the on-call engineer sees it immediately.
  • Automated deploy visibility. When deployments happen automatically (for example through GitOps), notification rules surface failures of preview and Git push deploys.
  • Incident paging. Send incident opened, escalated and resolved events to the channel your on-call team watches.
  • Multi-team coordination. Scope rules per project so each team receives only its own deployment events.

When NOT to use notifications

If you only deploy occasionally and watch the CLI output yourself, you may not need notification rules. They become important once deployments happen automatically or across many projects and stages.

Notifications are also not the tool for runtime health monitoring. To alert when error rates spike, CPU saturates or latency crosses a threshold, use alarms, which watch CloudWatch metrics on running resources.

Event types

Each rule lists the events it watches. The New notification rule dialog groups them as follows:

GroupEvents
DeploymentDeploy started, Deploy succeeded, Deploy failed, Deploy rolled back
DeletionDelete started, Delete succeeded, Delete failed
RollbackRollback succeeded, Rollback failed
Git deploymentsPreview deploy succeeded, Preview deploy failed, Git push deploy succeeded, Git push deploy failed
ScriptsScript run succeeded, Script run failed
SecretsSecret created, Secret updated, Secret deleted
SSM parametersParameter created, Parameter updated, Parameter deleted
IncidentsIncident opened, Incident escalated, Incident resolved
SecurityNew critical or high security findings

A new rule starts with the failure events, the incident events and new security findings selected. A single rule can watch several event types, so related alerts can go to one channel through one rule.

Managing notification rules

Open Channels in the Alerting group of the Stacktape Console and click a channel. Its page lists what the channel receives, one rule per row:

ColumnDescription
EventsThe events the rule sends, grouped by subject, for example "Incidents: opened, resolved"
FromThe projects and stages the rule covers, for example "All projects · production stages"
AddedWhen the rule was created

The Channels list summarises the same thing per channel in its Receives column.

To add a rule, click New rule on the channel's page, choose the event types and, optionally, projects and stages, then click Add rule. To delete a rule, use Delete in its row menu; the channel stays. Adding and deleting rules needs the org:manage-notifications permission (organization owners and admins); every member can see them.

A new Slack (Stacktape app) channel starts with default rules: incidents and failed deployments from every stage, successful deployments from production stages, and new security findings.

Scoping notifications

Rules can be scoped to specific projects and stages. When a scope is left at its default, the page shows "All projects" or "All stages".

  • Project scoping restricts the rule to specific projects. Only events from those projects send an alert.
  • Stage scoping restricts the rule to specific stages, for example only production. This keeps dev and staging deploys from generating noise.

Project and stage scoping can be combined. Secret and SSM parameter events belong to the whole organization, so a rule that watches only those events has no project or stage scope.

Create separate rules for different severity levels. For example, route failure events to a high-urgency Slack channel and success events to a low-noise webhook.

Checking what was sent

Sent alerts in the Alerting group lists every alert Stacktape sent to your channels, with the delivery outcome per channel. Each channel's page also shows its recent alerts. If an alert was not delivered, the delivery details name the error and link to the channel, where Send test checks it after you fix its settings. The old /notifications address opens Channels.

Notifications vs. alarms vs. budgets

Notifications, alarms and budgets serve different purposes. Notification rules and budgets send to Console alert channels; an alarm sends to the notificationChannels listed on it in the configuration and opens an incident when it fires.

FeatureWatches forDefined in
Notification rulesStacktape events (for example a failed deployment or an opened incident)Console, on each channel's page
AlarmsCloudWatch metric thresholds (error rates, CPU, latency, queue depth)Stacktape configuration, alarms list on a resource; created on deploy
BudgetsAWS spending thresholdsConsole, Budgets page

Use notifications for operational awareness ("did the deploy finish?"), alarms for runtime health ("is the API throwing errors?") and budgets for cost control ("are we overspending?").

Stacktape creates the AWS budget when a budget is saved. Alarms are created when the stack that defines them is deployed.

Notification Rules FAQ

Can I route different events to different channels?

Yes. Each channel has its own rules, so give each channel the events and scope it should receive. For example, send all deployment events to a low-noise webhook with one rule, and route only production deploy failures to a high-urgency Slack channel with another.

Why didn't my notification get delivered?

Open Sent alerts, which lists every alert that was sent with its delivery status per channel. If a delivery failed, the cause is usually the destination alert channel: an expired webhook URL or a misconfigured Slack integration.

What is the difference between notifications and alarms?

Notifications watch for events that Stacktape records, such as a deployment failing or an incident opening. Alarms watch CloudWatch metrics on running resources, such as CPU usage, error rates, latency or queue depth.

Do I need a separate alert channel for notifications and budgets?

No. The same alert channel can be used by notification rules and budgets. To get alarm alerts in that channel, add a rule that watches incident events: a firing alarm opens an incident.

Can I get notified only for production deploys?

Yes. Set the rule's stages to the stage names it applies to, such as production. Combine it with project scoping when only some projects should send alerts.