Learn about performance testing strategies for an ecommerce platform.
I struggled to find useful resources here.
But, after a bit of reading and scanning various web sources, here’s something I learned.
It’s really tempting to stress test a happy path.
Search for an item, add it to cart, complete purchase.
You might think that setting up this path, and then ramping up a large number of runs in a tool like JMeter, might be enough.
But - that might not be true!
It’s possible that a search is a much more processor intensive process than adding an item to a cart.
It’s also possible that a search is done much more frequently than a purchase - there are possibly lots of people searching but not buying.
This throws the performance test out of balance - it’s no longer a realistic test of performance.
Performance tests need to account for real user behaviour, and functions that are more processor intensive than others.
Ideally, some numbers on user behaviour can be gleaned beforehand.
Otherwise, start thinking outside the happy path, before designing your performance tests.
Hope that makes sense! Sorry that’s a short one!