Define the slow experience
Is initial page load slow, are interactions delayed, does checkout pause, or is the backend/admin the problem? Record page types, geography, device and time patterns.
- Frontend pages
- API requests
- Admin/backend
- Background jobs
Measure the browser and network
Look at document response time, resource waterfall, JavaScript execution, images, fonts, third-party scripts and rendering milestones.
- TTFB
- Resource waterfall
- Main-thread work
- LCP/INP/CLS
Profile the application
Backend time can come from database queries, template rendering, external services, search, sessions, cache misses or custom code. Use logs and profiling rather than assumptions.
- Slow queries
- Cache hit/miss behaviour
- External API latency
- Application errors
Check infrastructure in context
CPU, memory, disk, PHP workers, database resources, search and cache services matter, but infrastructure should be correlated with application behaviour and traffic.
- Resource saturation
- Worker limits
- Database/search health
- Scheduled jobs
Questions to consider before you start
Should we upgrade the server first?
Only if evidence points to resource constraints. Bigger infrastructure can hide inefficient code without solving the underlying issue.
Are PageSpeed scores enough?
No. They are useful indicators but do not describe backend, admin, integration or intermittent production performance.
How long should an audit run?
Some issues are visible quickly; intermittent problems may require monitoring across representative trading periods.
Can you perform the audit for us?
Yes. Midoriweb provides performance review and implementation work across Magento and wider web platforms.
Explore connected services
Need a performance investigation?
Send the URL, the slow user journey and any existing evidence. We can help establish a useful diagnostic scope.