Thanks so much to @gregraven for talking through this with @muhzulzidan ! Greg gives a lot of the same advice our staff would:
- make sure you understand what is being tested
- allow for things to be different locally than on the network
For our staff to debug further, we’d need to see specific slow requests. If you’d like to capture an x-nf-request-id HTTP response header value (see this article for how to do it), then we can look at our internal logs and correlate the response.
Even better would probably be a HAR file (see HAR Analyzer for details about collecting it), since that will show all the interactions of your page. One from gatsby cloud and an identical URL from Netlify would be the best side-by-side comparison setting and I won’t be shy about admitting it, in case their service looks better than ours 