Boost up your freaking servers Kabam
UniversalGaming
Member Posts: 723 ★★★
I've a Wifi with speed about 100Mbps
Even then I'm getting the problem connecting to server error
Boost up your server for god's sake!
You earn millions from the players yet can't upgrade your servers
Even then I'm getting the problem connecting to server error
Boost up your server for god's sake!
You earn millions from the players yet can't upgrade your servers
Post edited by Kabam Zibiit on
40
Comments
Man, asking for help in the arena is like a full time job with those servers
This explains how connections work.
https://www.google.com/amp/s/www.dnsstuff.com/reduce-packet-loss/amp
This touches on why it won't connect while still having bars.
While all that's happening, people are still playing arena, AQ, AW, story or just being logged in. All that gameplay is tracked live on the server.
With all those people playing, latency is something that can cause issues as well. If you have a high latency, you're going to see lag in help requests or sending help. Any game out there that plays on a live server has these issues. You can see examples of it everywhere.
The loading screen takes a while because the game is making it's initial connection to the server. That connection could be congested, you connection might be dirty or many other reasons.
Sometimes it's Kabams side, most of the time it's outside.
But 1000 loyalty a day? And a few mins off your cooldown?
Is that really worth the server issues, lags, and crashes it can cause?
Another "waaah your servers suck" post without any actual info or knowledge.
Moving along....
Maybe they think the 1000 loyalty is worth it to keep it in game.
So coming back to the discussion, yes we all see the loading circle sometimes and there can be many reasons for it, not just kabam servers.
1) Kabam isn't a server hosting company, they rent servers and decide how much of it they want.
2) Your ISP could be routing the request through multiple points rather than taking the shortest available path to the server. Any of these points could be choking the request-response.
3) If you have too many users on your home network that'll reduce the response time as well and show the loading circle.
4) There are home devices that interfere with signals too like baby monitors and other electronics. Your house can have wifi dead/weak zones as well, sometimes even next to the router.
5) Similar concepts apply to your carrier signal. You might be seeing full bar connectivity but there could be packet loss in the background resulting in this loading circle.
6) You cellular carrier could be overextended during peak hours or they have too many subscribers in a small area, leading to congestion.
So yes, it's easy to blame kabam for everything but it's not that simple. They can enhance their netcode as much as possible but it can never account for all the factors mentioned above. It's a compromise between performance and price.
http://imgur.com/gallery/y6KmX9n
So yes, I might be an exception, but consider the fact that you or your ISP or cellular carrier might have problems too.
During the best hours, I've even logged in within 12 seconds.
Alliance chat, Reddit, YT and the forums here all point to connection issues. You could argue that that only represents the <1% vocal minority, however, I doubt that to be true.
All I am pointing out is that it isn't ALWAYS the servers Kabam uses, and its very possible someone's connection could be the cause.
I agree it isn't always Kabam's fault, doesn't mean that it often isn't. Google don't necessarily help you architect your infrastructure. They just provide you the tools. Kabam might not be auto-scaling or they might not be using appropriate thresholds to trigger scaling... or they could be capping to prevent their costs blowing out. They might not be scaling enough new instances or they might be scaling with the wrong types of instances.
Google share or wholly own a significant portion of internet backbone. They don't have connection issues, which means that it's likely to do with what is happening inside Kabam's provisioned infrastructure.
Also game systems can’t necessarily be simply scaled up by spinning up more instances or something like that. The data back end, for example, that is probably the performance limiter for things like help requests, crystal openings, or sig stone usage, is probably transactional or otherwise atomic locked. Those can’t be arbitrarily scaled without being redesigned from the ground up.
Also, if the problem was simply a lack of compute resources, the problems players were experiencing would be totally different from what we’re actually experiencing. Literally running out of CPU or something similar has a very specific signature, and it is one I don’t see. We should see more of a time dependency, because player demand should fluctuate over a 24 hour period, and players playing during off hours should see less issues. This doesn’t seem to be the case. We should see certain activities more impacted than others, because we should see activities partitioned on compute resources most heavily utilized affected more than others, but we don’t see activity partitioning either. And we should see more cascade failures when running out of resources like this, and such failures also seem to be not commonly observed. While the game could be running out of server side resources as a part of the problems experienced, it can’t be a primary cause of the problems.
Performance scaling is one of those things everyone assumes is relatively straight forward, when in fact it is one of the black arts of information systems design. Every performance issue that someone believes was solved with a relatively simple solution like “just spin up more servers” was actually not a real performance problem.
And no, I never said just spinning up more servers solves the problem. Spinning up the wrong type of instances won't necessarily help either. You have memory optimized, compute optimized and general purpose. Ephemeral storage on SSD or HDD blah blah blah. A whole host of difference resources that can be scaled in a number of different ways.
You have load balancing, you have read replicas, you have sharding etc etc. Let's not forget about multiple availability zones and regions.
Yes, I get it, it is complicated... but it's not a dark art. Why do you think there are certifications in cloud engineering? They're not a liberal arts degree.
You don't see a time dependent signature, because that demand should be somewhat predictable. What I do see are issues related to demand, which is why I specifically mentioned content drops. I've seen so many times YouTubers have a stream waiting as the new content drops and the game is laggy as hell. Why? Because the infrastructure is hit with a spike of activity. The whole point of auto scaling is to provision according to demand and not to any peak or average threshold.