Comment

Somehow a multibillion dollar company can’t compete with an ancient quad core server shoved in a coat closet when it comes to page load times.

To be fair, it’s nearly impossible for remote sites to beat on-prem page load times, given the added per-component transit times over the internet.

Sort:hotnew top

Powderhorn@beehaw.org ⁨16⁩ ⁨hours⁩ ago
Shhh … the cloud is the only way forward. Heresy!

source
nbailey@lemmy.ca ⁨16⁩ ⁨hours⁩ ago
Totally true, but I’m talking an order of magnitude or two difference…

A query returning ~100 jira tickets would take about 250-300ms on our old beater running Postgres on busted old SAS drives shared with a bunch of other crap. Seek times were atrocious but not catastrophic. It usually didn’t timeout, and only crashed once in a while.

Sunning the same search on jira cloud now takes 2-3 seconds, often even more because the page first has to load 20 MB of JavaScript bullshit. Time from clicking a link to seeing information is so long you’ve got enough time to take a sip and put the coffee down.

Like I get it, distributed systems are hard. And having a multi tenant system as big as they run is probably crazy complicated. But come on, there’s no excuse for that level of consistently bad performance!!

source