Skip to main content
At 18:47 monitoring raises dozens of subscriber alarms and two core BGP-session-down alerts. The support agent’s assignment spans the entire incident: establish the affected services, collect evidence, coordinate the carrier and your engineers, keep subscribers informed, and verify recovery before closing the case. The cloud worker handles the actions its configured tools support. Engineers use Studio for router inspection and other local work; an unavailable carrier or communication action becomes an explicit human task. The support loop connects those handoffs so the incident continues toward restoration.

​
Workflow dependencies

Named systems are example dependencies. Verify cloud tool support and credentials before assigning actions. A worker does not inherit the desktop’s router SSH sessions or access to the carrier’s browser portal. Connector writes require an independent authorized manager to approve the exact request, including external updates, customer replies, and ticket closure. A workflow stage marked automatic does not bypass that approval.

​
From alarm to restoration

1

Correlate the incident

Retrieve the monitoring events through the configured connection. Group likely downstream symptoms under the candidate upstream incident while preserving unrelated customer faults as separate work. Use eligible Org Brain knowledge to find prior response guidance, then verify it against current evidence.
2

Request and assess live routing evidence

Assign an engineer a human task to inspect both core routers and return BGP state, neighbor reset reasons, relevant timestamps, and subscriber-path checks. Compare that evidence with available looking-glass data. Record what is confirmed and what still requires the carrier’s investigation.
3

Open and track the carrier case

Prepare the carrier request with peer addresses, ASNs, UTC timestamps, relevant logs, and an approved callback contact. Submit through a supported, authorized integration or an assigned person. Record the case reference and who owns the next carrier follow-up.
4

Establish subscriber impact

Query the configured subscriber system to identify affected services and regions. Validate the contact list and proposed message before sending. Include the known scope, current status, and next update point without inventing a restoration estimate.
5

Coordinate action and customer updates

Route consequential network decisions to the responsible engineer. Use available integrations for approved customer updates, or create human tasks for SMS, IVR, or status-page actions that the worker cannot perform. Keep the communication record linked to the incident rather than assuming one channel automatically updates the others.
6

Follow the incident through recovery

Configure an explicit follow-up workflow to recheck monitoring and carrier progress. Have engineers carry out approved infrastructure work and return the resulting evidence. If an approval, carrier response, or repair is outstanding, retain a named next owner and escalation path.
7

Verify service and close the loop

Require healthy routing plus appropriate subscriber-path or service tests. Prepare the all-clear from those results and update every communication path used during the outage. Close the carrier and internal cases as appropriate; keep unresolved subscriber faults open with their own owners.
8

Capture the incident lesson

Retain the timeline, actions, validation, and unresolved follow-up in the incident record. Save reusable knowledge with an explicit audience and review the worker’s outcome before revising its instructions.

​
What success looks like

  • Alarms become an actionable incident with evidence and an accountable response.
  • Carrier, engineering, and communication tasks remain part of the same support outcome.
  • Customers receive updates through routes that have an assigned owner.
  • The all-clear is based on restored service, with remaining faults carried forward explicitly.

Digital workers

Configure triggers, cloud tools, human tasks, and outcome review.

Procedures

Give operators a repeatable routing investigation and validation playbook.