Data Flows — MITx Pro

Generated 2026-06-24 16:33 UTC · c4gen dev

Each scenario below replays one interaction as a C4 Dynamic diagram. Amber steps are asynchronous (queued / scheduled / event-driven).

Paid enrollment & checkout (synchronous)

A learner pays through CyberSource Secure Acceptance. xPRO signs a payload and redirects the browser to CyberSource; CyberSource posts a signed result back to xPRO's OrderFulfillmentView, which fulfills the order, enrolls the learner in Open edX, syncs the deal to HubSpot, and emails a receipt.

Open edX SSO & user provisioning (synchronous)

xPRO authenticates learners through Open edX via social-auth. On first login/signup xPRO creates the corresponding edX user and an OpenEdxApiAuth access token, used later for enrollment and grade reads.

Courseware sync & certificates (asynchronous)

RedBeat-scheduled Celery tasks repair failed enrollments and faulty edX users, sync course-run data and grades from Open edX, and generate course certificates. External vendor courses (Emeritus / Global Alumni) are synced from report APIs on a daily cron.

Google Sheets coupon/refund/deferral ops (asynchronous)

Staff manage coupon assignment, refund, and deferral requests in Google Sheets. Drive push notifications and a polling beat task make Celery read the sheets via a service account, apply coupons/enrollments, email bulk enrollment codes, and write status back.

MIT Learn catalog ingestion (cross-service, asynchronous)

MIT Learn's ETL pulls xPRO's course/program catalog from the public REST API and content files from S3 to surface xPRO offerings in discovery. This is a one-directional pull owned by MIT Learn's Celery scheduler.

Ingestion sources (ETL)

Every external source the edx_content / default Celery workers pull from, with transport and cadence. ⚠️ marks brittle linkages (HTML/token scrapes, hardcoded URLs).