Black Friday, television campaigns or suddenly popular offerings not only change the number of visitors, but also their behavior. Many people search for the same products at the same time, use filters and expect current inventory.
Use a traffic mix with autocomplete, searches, categories, filters and product clicks. Take popular and heavy search questions and test multiple stores when they share the same infrastructure. A test with one simple request gives a misleading picture.
Check performance after a new rollout, scale change, or blank cache. Measure what happens when many visitors search for the same and different products at the same time. Caches are allowed to improve speed, but do not continue to show an outdated price or stock.
Determine which additional functions may be temporarily easier when response time increases. Core features such as search and product navigation take precedence over optional enrichment. Test the transition in advance, so that simplification during a peak does not cause a new malfunction.
Designate those responsible, choose one shared overview and set escalation boundaries. Reduce risky changes just before and during the peak. Check afterwards whether capacity, alerts and playbooks met and process the findings for the next campaign.
Discuss which measurements, responsibilities and restoration routes fit your online stores and technical environment.