Configuring Identity-Scoped Telemetry Alerts
Scouts often operate within a defined scope — a specific jurisdiction, a specific business unit, a specific persona pool. The alerting surface should respect that scope: a Scout responsible for EU customer support operations should be alerted when EU support metrics breach thresholds, not when US sales metrics do, and not when a global system event fires.
Lookout's identity-native alerting handles this without you writing filter conditions. The alert engine pre-scopes results to your hardware-attested identity context — your jurisdiction, your persona pool, your operational scope — so the alerts you receive are relevant by construction.
This tutorial walks through configuring a set of identity-scoped alerts for an EU customer-support Scout.
Prerequisites
- A Scout-role account with
subject.attributes.jurisdiction = EUandsubject.attributes.persona_pool = customer-support - Lookout installed
- Familiarity with one or two dashboards you already watch operationally
Step 1 — Understand the identity context
Open the Console's profile menu and click My identity context. You should see:
Subject: usr_<your_id>
Type: user
Roles: [scout, customer-support-lead]
Attributes:
jurisdiction: EU
persona_pool: customer-support
operational_region: [DE, FR, IT, ES, NL]
MFA: verified at <recent timestamp>
Gate attestation: chain <id> (verified)
These attribute claims are what Lookout will scope alerts against. If any of these do not match what you expect, surface to your Pathmaker — they govern subject-attribute provisioning.
Step 2 — Configure your first scoped alert
In Lookout, navigate to Alerts → New alert. Configure:
| Field | Value |
|---|---|
| Alert name | support-queue-spike-EU |
| Metric | gauge(support_tickets.open) |
| Filter | [automatic: identity-scoped] |
| Condition | > 50 |
| Window | 5 minutes sustained |
| Channel | in-app + email to your address |
| Cooldown | 30 minutes after resolution |
The [automatic: identity-scoped] filter is the key feature. When set to automatic, Lookout applies your subject.attributes.* claims to the metric query. For this support metric, the result is automatically filtered to tickets in EU operational regions you have visibility into.
You can inspect what the automatic filter expands to by clicking Show expanded filter:
filter:
and:
- eq: ["resource.jurisdiction", "EU"]
- in: ["resource.region", ["DE", "FR", "IT", "ES", "NL"]]
- eq: ["resource.persona_pool", "customer-support"]
The expansion is the union of constraints from your subject attributes that the metric's resource schema can bind to. You did not write the filter; Lookout assembled it from your identity context and the metric's resource declaration.
Step 3 — Add a "first response time" alert
Alert: support-first-response-slow-EU
Metric: histogram(support_ticket.first_response_seconds)
Filter: [automatic: identity-scoped]
Condition: p95 > 300
Window: 1 hour
Channel: in-app
Cooldown: 1 hour
The 95th-percentile first-response time is under 5 minutes (95% of tickets are answered faster); this alert fires if that degrades for an hour. Scoped to EU support tickets only — a US first-response slowdown does not bother you.
Step 4 — Add a sentinel for cross-scope leakage
If you ever receive an alert about a metric outside your scope (an alert about US tickets reaching you), it's a configuration error. Set up a sentinel:
Alert: scope-leak-sentinel
Metric: count(alert_deliveries)
Filter:
and:
- eq: ["alert_subject", "usr_<your_id>"]
- or:
- neq: ["alert_resource.jurisdiction", "EU"]
- and:
- exists: "alert_resource.persona_pool"
- neq: ["alert_resource.persona_pool", "customer-support"]
Condition: > 0
Window: instant
Channel: in-app (high priority)
If any alert is delivered to you whose resource is outside EU customer-support, the sentinel fires. The sentinel is your safety net against misconfigured automatic filters.
Step 5 — Test the alerts
Lookout's alert tester injects synthetic events to verify the alert fires correctly. For each alert you configured:
- Click Test → Generate synthetic event
- Confirm the alert fires
- Confirm the notification arrives in your in-app inbox / email
For the sentinel, generate a synthetic alert with a US-scoped resource and confirm the sentinel catches it.
Step 6 — Group alerts into focus modes
Lookout supports grouping alerts into focus modes — sets you can toggle on/off depending on whether you're actively monitoring or off-duty:
Focus mode: EU-support-on-call
Alerts in scope:
- support-queue-spike-EU
- support-first-response-slow-EU
- scope-leak-sentinel
Notification preferences:
- In-app: always
- Email: when ON
- SMS: when ON + condition severity is "critical"
When you're on-call, toggle the focus mode on. When you go off-call, toggle off — alerts still log in the in-app inbox but stop interrupting via email/SMS. The handoff to the next Scout on call happens cleanly.
Step 7 — Observe alert delivery in the audit log
Every alert delivery is itself audit-logged. The platform records:
- The alert that fired
- The metric value that triggered it
- The subject who received the notification
- The delivery channel
- The acknowledgement (if you acted)
Wardens can audit alert delivery patterns: "did the right person get notified within the response SLA?" The answer is in the chain.
What just happened
You configured a set of identity-scoped alerts that respect your operational boundary. The benefits:
- No false-context alerts — you do not get notified about regions or persona pools outside your responsibility
- Structural scoping — the scope is your hardware-attested identity, not a filter you can forget to update
- Sentinel coverage — a misconfiguration that would leak cross-scope alerts to you is itself surfaced
- Focus modes for on-call discipline — the alerting surface matches your operational state
- Audit-logged delivery — the alert chain is forensically reviewable
Next
- Building a Governed Business Workflow — connect operational flows the alerts watch.
- Generating Audit-Ready Operational Reports — pull periodic reports from the same metrics.
- Pathmaker → Identity-Native Telemetry Mapping — the dashboard surface the alerts are derived from.