I joined Lemmy a few days ago under the lemmy.world instance and want to keep it as my main instance, but it’s being pretty laggy.
I don’t have access to a computer to ping each instance so am wondering if there’s a mobile way to do so.
The lag is a result of server load, not ping. Lemmy.world is the largest lemmy server and it’s groaning under the weight of the July 1st refugees.
It’ll resolve itself as servers get upgraded or people spread out. You can try joining another smaller instance if you don’t want to wait.
That gave me an image. Think of Atlas, the guy who had to hold up the world, only replace that person with the individuals trying to keep up with server demand of lemmy.world
If I could do anything in artwork or Photoshop I’d create a meme myself.
“Ping” shouldn’t affect lag, and your ping should be almost unnoticeable to you. Ping basically means the time a packet takes to the server.
For anything other than a video game, you’re going to focus more on server load, and utilization. If something is overloaded (Lemmy in general is overloaded) it doesn’t matter what your ping is, the server can’t process as many requests as it gets and you see “Lag”.
Part of the process right now is figuring out if Lemmy can keep up with the traffic, this is part of the growing pains of a new “social media” server. Best thing to do if it’s bothering you is give it a few days and hope they can work something out.
I realize. I just wanted to make a new account with the lowest ping instance while the server is upgraded at lemmy.world
The point I’m making is “ping doesn’t matter”. If you want to go to a different instance that’s fine, find one that seems responsive. You’re using the wrong terminology/thinking when deciding which instance you want to be on.
It doesn’t? Something with 1000ms is going to be less responsive than 100ms geographically closer, though I don’t know how much it varies.
I assume that me, being the the US, will have a significant enough high ping by being on an instance that’s hosting in Finland. I never used the wrong terminology, just was asking about how to find the lowest ping — I only said I was trying to find the lowest ping because lemmy.world’s servers are shutting the bed. I can see where the extrapolation comes from though, didn’t make it clear enough.
Unless you’re clicking through pages at record speed, 1 second latency for a web page and 100ms will be almost unnoticeable.
I just pulled Amazon.com, and on my computer it takes about 2 seconds for it to render, if we add 1 second of ping and it took 3 seconds to rend, I wouldn’t notice it. Looking inside chrome tools, I see it takes 700ms to download the content on the front page. This is relative of course and I could go deeper (if the initial page, and the content server both were 1 second away, technically it could take 2 seconds because it needs two round trips) but that’s kind of besides the point.
The main issue on the web is content generation/download, it’s not the time to reach the server. Lag matters more in gaming because you’re constantly talking to the server in a round trip constantly, so any latency is increased and will be more problematic. but with HTTP, you’re sending a request and getting it back, it’s a single round trip, you then will take time to parse that data and read it, and then when you’re ready for more you’re doing another round trip.
So if a server is 100ms or 1 second away you’re only paying that penalty once in a while. The issue is if the website gets under heavy load and can’t respond for multiple seconds or more, or fails to respond. That’s more of an issue with server load, which is why I say to focus on that.
If you TRULY think this is an issue, go to your command line and run “ping lemmy.world” I get 189ms pings. Amazon gives me 79 ms pings. Google is 28 ms ping. Those levels will be unperceivable after you consider the rest of the time it takes to download and render a page.
The difference between 2s and 3s is actually quite large, in terms of peoples patience to stay on a website. There have been many studies on the effects of longer RTT for websites. The conclusion of most of these studies is that there are massive drop-offs in users (abandoned sessions) once you get into the 3, 4, 5s ranges.
The difference betwen 1000ms and 100ms might be massive for gaming or live video chat as that 1 second delay is very noticeable. In terms of viewing a web page, a second to send a request from click will be almost imperceptible. As there is still a few more seconds to load the data that your device anyway. So you are talking about a 5 second wait time or a 4 second wait time. It doesn’t really matter.
The big difference is server load. A quiet server with a fast internet connection doesn’t need to process as many requests and therefore you don’t have to wait in queue. A low server load with a high ping could take a second or two while a server next door to you with 10 seconds worth of requests could take ten seconds.
The ping isn’t the issue as real time delay is not the important factor here.
Try a few instances and see which feel faster. It’s kind of hard to guage, although I imagine someone could build a tool that tests server responsiveness at pulling a request from a sample link and a local link on its own instance and generate a report I guess.
There is unfortunately not a good tool to help with what you’re looking for.
The best way is to browse to a few servers and see if they feel snappy. You can click through without an account.
https://wirebase.org is ours, which is hosted in the US
My understanding (from limited knowledge) is that also due to how federation works even if you’re instance isn’t under too much load, you may notice issues with posts/comments from other instances if they’re struggling.
Start with physical location https://lemmy.fediverse.observer/map
then ping/test the ones you like closest to you.
It’s weird. My instance is hosted on a VPS in New Jersey, but shows here in Mexico. The company I rent the server from doesn’t have a data center in Mexico
they providing any edge caching for you?
Not as far as I can tell. It’s a pretty barebones VPS from IONOS
somehow this is dns’s fault
Yea ended up choosing a pretty popular one near me, hopefully they don’t shut down or something though. Thanks!
always a risk. keep an eye out for migration tools.
I dont have it on me but there are fediverse sites that track all the instances.
I’d like to see or make a tool that measures “federation” quality.
Lots of servers are locally responsive, but lag or completely fail posting to remote servers or accepting remote requests.
How can i participate in a conversation when half of the instances are hours behind or may not get my comments at all :(
This seems to be a combination of software and load and it is very insidious because it doesn’t affect local functionality so admins that don’t care or check think their communities are just fine.
I’ve been considering doing something very similar. Will PM you
