Customer removed a proxy in front of @vercel and saw a 50% reduction in TTFB. What's cool about this view is that it's Speed Insights, which grabs data from end users devices. It's as close as you can get to the truth. Generally speaking, developers get too caught up on "how it feels on my machine" or "I saw some random benchmark on the internet". I encourage you to always look at the field data to draw conclusions. Hence why this is called "RUM" (Real User Monitoring). Any RUM tool should work (@getsentry, @datadoghq, etc). Source: https://t.co/4sBMW0LKZ6
