Skip to content

Network and IT infrastructure use case

Investigate VPN tunnel failures and branch connectivity

A tunnel that is down, repeatedly reconnecting or carrying no useful traffic needs a scoped investigation. NetPrompt AI can help organize a VPN troubleshooting workflow using your network context, approved device checks and connected monitoring evidence.

Define what is failing

Start with the affected sites, VPN type, peer identities and time window. Distinguish a negotiation failure from a tunnel that appears established but cannot carry the required application flow. Confirm which prefixes, routing context and services are expected to use it.

Describe the expected primary and alternate paths in Network Context. Troubleshooting Context should specify the checks appropriate to the deployed vendor and VPN type. A diagnostic procedure for one implementation should not be assumed to work unchanged on another.

Compare transport, tunnel and service evidence

Organize observations by layer and collect them over a comparable time window. Cisco’s troubleshooting guidance highlights the value of comparing the underlying transport with the tunnel during an intermittent event.

Compare transport, tunnel and service evidence
LayerUseful evidenceQuestion
Underlying transportPeer reachability and path observationsWas the transport available during the failure?
TunnelRelevant negotiation state, logs and traffic countersDid the tunnel establish and exchange expected traffic?
Routing and policyAffected prefixes, routing context and security rulesWas the intended service flow permitted and routed?
ApplicationRepresentative transaction from the affected siteDid the service work through the expected path?

Illustrative example: established tunnel, failed application

A branch reports that an internal service is unavailable although the VPN status shows established. The workflow collects the approved tunnel, routing and policy observations for the affected flow. The engineer checks both directions and compares them with a working service.

The result should explain which evidence supports the suspected fault and what remains untested. Tunnel status alone does not prove end-to-end service availability. This is an illustrative evaluation scenario, not a customer result.

Configure the NetPrompt investigation

Use supported device access and reviewed command templates to collect the required observations. Add monitoring or API evidence where available, and verify platform coverage before running the workflow. NetPrompt’s general orchestration capability does not imply a prebuilt VPN procedure for every vendor.

Begin with evidence collection. A tunnel reset or policy change can disrupt other flows and should follow the applicable approval process. After an approved action, repeat the service test and retain the before-and-after observations.

  • Record sites, peers, VPN type and affected service.
  • Confirm template coverage and target identity.
  • Align observations to the incident time window.
  • Separate transport, tunnel and application findings.
  • Review any proposed action and verify service recovery.

Test intermittent and incomplete cases

Use one known fault, one healthy service and a case where peer-side evidence is unavailable. The workflow should identify the coverage limit instead of inventing the remote state. Measure the usefulness of the explanation and remaining manual investigation against your baseline.

Common questions

Does an established VPN prove that applications work?
No. Routing, policy, return traffic and application behavior still need verification for the affected flow.
Does NetPrompt support every VPN vendor automatically?
Validate the device profile, software version, access and diagnostic commands for the chosen VPN. This page describes a configurable investigation, not universal turnkey coverage.
Should the workflow reset the tunnel first?
Start with evidence. Disruptive actions require an appropriate scope, approval and recovery verification.

Continue exploring

Further technical reading

External references explain general practices; they are not endorsements of NetPrompt AI.

Evaluate the workflow on your network

Bring a representative incident, your device and OS versions, and the tools your team uses. Review evidence quality, integration scope and approval controls in a technical demo.