While running perfomance tests I noticed that there is additional wait time across the board now compared to 2 months ago.
As an example, our favico download time increased from 16ms to 50-60ms, exclusively due to wait times, in both instances the response header included:
cache-control: public, max-age=86400000, s-maxage=86400000, stale-while-revalidate=86400000
cf-cache-status: HIT
While the individual contribution can be considered minor, it translates into a significant impact to performance metrics for the page as a whole.
I would be grateful for any suggestions regarding:
How to debug / pinpoint the source of the additional overhead.
Thank you for the suggestion, it looks like a nice tool - unfortunately the values there do not match what I see from GTmetrics, using a fixed location ( HK ) nor running locally, in both cases the TTFB are much higher, even when running multiple times in quick succession against page resources that identify as cache-hit and controlling for the region the requests are made from.