Observability on Oxygen: Monitoring Shopify Edge Apps

Introduction

Hydrogen apps running on Oxygen are fast and global — but they’re also opaque if you don’t set up observability. Logs, metrics, and traces aren’t optional; they’re critical for performance, debugging, and client trust.

This post shows how to monitor Oxygen storefronts effectively, from built-in tools to enterprise integrations.

Why Observability Matters

• ⚡ Performance budgets → catch TTFB spikes early.
• 🐛 Debugging edge failures → Oxygen errors often don’t reproduce locally.
• 📊 Client reporting → agencies need proof of uptime + speed.
• 🔒 Compliance → some industries require log retention + audits.

👉 Observability = visibility into the edge.

Built-In Options

• Shopify Oxygen provides profiler logs.
• Basic metrics: request latency, error rates.
• Limitations: non-Plus merchants get limited retention + detail.

Plus Merchant Benefits

• Log drains → forward logs to Datadog, Splunk, New Relic.
• Full-stack observability (logs + metrics + traces).
• Enterprise SLA compliance.

Non-Plus Merchant Workarounds

• Implement lightweight observability in Hydrogen itself:
  • Add structured logging (JSON logs).
  • Use third-party services like Logtail or Tinybird.
  • Track RUM (real user monitoring) with small beacons.

console.log(JSON.stringify({ level: "info", route: request.url, ttfb: response.headers.get("x-oxygen-ttfb"), }));

Case Example: Agency Setup

• Client wasn’t Plus → no Datadog integration.
• Agency added structured logs + RUM pipeline.
• Collected performance + error data in BigQuery.
• Outcome: agency delivered monthly “Oxygen Health Reports.”
• Result: upsold client to recurring monitoring retainer.

Guardrails

• ✅ Always add structured logs in Hydrogen.
• ✅ For Plus merchants, set up log drains immediately.
• ✅ For non-Plus, layer lightweight observability.
• ✅ Use logs to feed performance audits + SLA reports.

Conclusion

Observability is the difference between flying blind and running a measurable, reliable Hydrogen storefront. Agencies should set up logging + monitoring as a core service, not an afterthought.

If you can’t see it, you can’t trust it.