How to investigate ecommerce problems before changing code
Logs, New Relic, CDN data, HAR files, PageSpeed, analytics and code inspection together give a far better diagnosis than rewriting components straight away.
Logs, New Relic, CDN data, HAR files, PageSpeed, analytics and code inspection together give a far better diagnosis than rewriting components straight away.
Across many incidents, combining production logs, New Relic, CDN data, HAR files, PageSpeed and code inspection gave a much stronger diagnosis than immediately rewriting components.
The evidence toolkit
- Application and server logs
- APM traces (New Relic)
- CDN and WAF logs (Fastly, Cloudflare)
- Browser HAR files and console errors
- PageSpeed and CrUX data
- Analytics funnels
- Code and config inspection
A simple process
Define the symptom, gather evidence, form a hypothesis, reproduce, fix the root cause and verify in production.
Common mistakes to avoid
How we help with development
Frequently asked questions
Why not just rewrite the component?
Rewrites often move the problem instead of fixing it.
How long does investigation take?
Usually hours, and it saves days of wasted work.
Can you investigate for us?
Yes — start with a store health check.