Provider Price Sync

Context

A digital-goods business sells products where a single SKU is often listed across 2–3 external supplier APIs. The storefront needs to always show the cheapest available price and route fulfillment to that supplier.

Problem

• Multiple supplier APIs with different shapes, rate limits, and auth
• One supplier ships an encrypted CDN price file (not an API), needs decrypt + cache
• Need to run frequently without breaking when a supplier is down
• Operational controls had to live outside the binary so we could enable / disable / rollback fast

What I built

A Go service deployed as Kubernetes CronJobs on EKS.

Architecture

• price-checker CronJob (every 5 min): batches through the catalog, fetches live prices from each enabled provider per product, picks cheapest, pushes via webhook to the storefront
• encrypted-feed-sync CronJob (every 30 min): downloads a supplier's encrypted CDN price file, OpenSSL-decrypts it, caches as a fast binary .gob
• Cursor-based batching: stores skip in a cron_job_settings table; auto-resets at end of catalog. Runs are idempotent and resumable.
• Operational controls: dry_run mode, per-provider enable/disable via JSON SERVICE_CONFIG Secret, CLI/env provider filtering, single-product debug mode for testing in prod without side effects
• Structured logging: log/slog JSON to stdout, parsed by Grafana/Loki with per-field queries (price changes, batch completions, errors)

Stack

• Go
• AWS EKS, K8s CronJobs (not system cron)
• OpenSSL for encrypted-feed decrypt
• Grafana + Loki for observability
• GitHub Actions → ECR → EKS

Outcome

Customers always see the cheapest price across suppliers. New suppliers can be added without redeploying. Flip a flag in the K8s Secret. Single-product debug mode means I can verify a fix in prod without batching the whole catalog.