API monitoring wired to PagerDuty: change validation time fell 85%

At UScellular in 2025, I built the WDH API Validation and Monitoring tool, a Python and Flask app. It ran API health checks on demand or on a schedule, in place of checks engineers had run by hand.

Summary

Role
Lead Continual Service Improvement Engineer, UScellular, Jan 2021 – Aug 2025. Enterprise CSI Lead for the Wing Digital Hub and FlowOne task force with Nokia. I built this tool in 2025.
Task force
A five-track get-well program, with an operational owner for each track on the DevSecOps platform team.
Results
Change validation time fell 85%. Several major outages were caught before customers noticed.
Tools
Python and Flask · REST and GraphQL health checks · PagerDuty · Postman, for the hand-run checks the app replaced

Problem

  • Engineers ran the API health checks by hand in Postman, several times a day.
  • Each run took about 30 minutes.

What I did

  • I built a Python and Flask app with a library of REST and GraphQL health checks.
  • Checks ran on demand or on a schedule.
  • A failed check opened a PagerDuty incident, and recovery closed it.
  • The app handled retries and token refresh.
  • It stored response times for trend and SLA reporting.

What changed

  • Running the checks took a few clicks.
  • Change validation time fell 85%.
  • Several major outages were caught before customers noticed.
  • Engineering leadership circulated the tool as an example of improving the network and the customer experience.

How a health check ran

  1. Scheduled or on-demand run
  2. REST or GraphQL health check
  3. PagerDuty incident: opened on failure, closed on recovery
  4. Stored response time
  5. Trend and SLA reporting
A check started on a schedule or on demand. The app ran the REST or GraphQL health check. If the check failed, the app opened a PagerDuty incident and closed it on recovery. Response times were stored for trend and SLA reporting.

Related case study: Service Impact Detection, an AI network analysis tool that went into production in the Network Operations Center.

If you're hiring for this kind of work, send a message.