Hello everyone, in my ubuntu server I am using Caddy to serve my sites, that’s working pretty fine. However, now I am trying to increase the speed of load of my site, I used Pingdom to analyze the site, and it gave me some curious information about the connection speed.
Distance from your server to Pingdom’s agent, connection quality, workload on your server, workload on Pingdom’s agent.
You have no control over Pingdom’s DNS resolver, so you can’t do anything about the biggest factor, unfortunately (unless your name server is really slow, but that’s probably not an issue for repeat requests).
SSL always introduces a short delay.
Have to admit some ignorance here as to what constitutes the “Connect” delay or why it’s that large. Here’s a test from Pingdom’s Melbourne agent to my Sydney web server, a Vultr VPS running Caddy:
Hey @Whitestrake thank you so much, it really was a complete answer. I will consider then that factors and see what can I get to improve it. Again, thanks a lot
I was checking thing by thing in my infraestructure, I already have another VM in the same server and the apps there respond really quick, the only difference that I found is that the slow server doesn’t have caddy installed as a service, instead it’s executing caddy as an app with ./caddy --conf caddyfile in other hand, the fast server has Caddy installed as a service, now am wondering: can be this one of the causes of so much time spent?
Can it be? Maybe, it’s not impossible that running it outside of a service manager is making it slower, and I’ve seen weirder things before. I don’t consider it likely at all, though. A service manager is more overhead (although negligible since it’s likely running on the host anyway, regardless of whether it’s overseeing Caddy as well).
