Skip to Content

Troubleshooting

Use this page when an early test does not match expectations and you need to understand where to look before changing configuration.

Start from visible operational signals: trunk state, Dialer Status, Real-Time Agents & Trunks, associated contacts, concurrent call capacity, and call records.

Change one layer at a time. Do not edit prompt, trunk, routing, and campaign settings together while you are diagnosing the same issue.

Advanced analysis with dashboards, logs, and traces

For advanced analysis, review the Grafana dashboards to understand whether the issue belongs to the system, telephony layer, database, container runtime, or workflow behavior.

Use Grafana Logs Drilldown when you need to move from a live symptom to the related logs and traces. This is useful when the problem crosses a campaign, API action, function call, endpoint, or internal service.

  • Dialer Dashboard shows outbound campaign status, active calls, limits, dialing rate, and live outcomes
  • Real-Time Agents & Trunks shows whether agents, endpoints, and trunks are available, in use, or unavailable
  • Logs Drilldown helps inspect logs and traces from Grafana when the issue needs deeper evidence
  • use Node Exporter Full to confirm host and system health
  • use Asterisk Overview to review telephony behavior at a higher level
  • use Asterisk Log when you need telephony-related log review
  • use cAdvisor when container runtime pressure may be involved
  • use MySQL 8 when the database may be contributing to the incident

If dashboards, logs, and traces do not show a system-level issue, continue with the inbound or outbound checks below and use the screenshots as the reference for what to verify.

What to check if something goes wrong

Inbound

If the inbound run does not behave correctly, start from Trunk status. If the trunk is not Registered, it cannot receive the test call and the inbound routing will not be reached.

  1. the trunk shows Registered in Trunk status
    Trunk Status screen showing registration Registered and endpoint Not in use.
  2. the inbound number or route is connected to the expected AI Inbound Routing
  3. the correct agent is assigned
  4. the agent has the right prompt
  5. the required functions are attached, if the workflow uses transfer, hangup, or API actions
  6. the result appears in Inbound Call Records

Outbound

For outbound, start from the campaign real-time dashboard:

  1. open the campaign real-time Dashboard
  2. check Dialer Status: if it shows Call in progress, the campaign is placing calls
    Dialer Status table showing Calls In Progress, Active campaign state, active calls, and call limit.
  3. check the number of generated calls and active calls
  4. check Real-Time Agents & Trunks: the trunk is Not In Use when available and becomes In Use while calls are running
    Real-Time Agents and Trunks dashboard showing endpoint states, Not In Use, and In Use timeline.
  5. check that the campaign has associated contacts and that Scheduled At is earlier than now
    Campaign contact result showing customer, phone number, availability, priority, and campaign identifiers.
  6. check concurrent call capacity, campaign start and end dates, active days, and timezone
    Edit campaign screen showing concurrent call capacity, active days, start and end dates, and timezone.
    Edit campaign screen showing concurrent call capacity, active days, start and end dates, and timezone.
  7. check the final result in Call Records

A practical recovery pattern

When something is unclear:

  1. return to one known agent
  2. use one controlled route or one small campaign
  3. place one manual test call
  4. review one call record

Then review the related operational signals:

  • active calls versus configured limits
  • campaign status and dialing activity
  • endpoint and trunk state
  • recent talk time and ring time trends

Next steps

Call RecordsReturn to the exact call record for the test you are diagnosing.Dialer Dashboard PanelsReview the outbound campaign panels for status, active calls, limits, outcomes, and live pressure.Monitoring and VisibilityReview the live operational layer around dashboards, logs, traces, telephony, and infrastructure.Logs and TracesUse OpenTelemetry-based logs and traces when the issue spans campaigns, endpoints, services, or API actions.
Last updated on
Quickstart