Server Performance issue for last few weeks

This is regarding the already discussed issue on FB:

Here is the link to one of the PUG where people complained about server performing poorly:
Hope, you guys keep demo’s with in-game voice recordings, otherwise it won’t be much helpful. I’ll try to find more PUGs like these.
This one might also have few complaints from players: (I’ll link this in comment, your forum doesn’t allow more than 2 links for new users)

Before you guys say that this my ISP’s or AWS networking issue, read the attachment.

Potential PUG where people complained about the server:

We are bringing up new IPs in Mumbai to see which is most prone to loss at the moment. We have 4 different ones at the moment. The games your are getting after midnight might be on a cluster that is working fine.

We can figure out and atleast have a way of fixing this till the ISPs sort their shit out.

So you can report the IPs where you are getting loss and where you arent.

I also did talk to our friends at LOT who host their community servers on AWS and they have reported the same. Their admin who lives in Mumbai is getting 40% loss to his own server.

Thank you for the the post @anX1 :slight_smile: As @akS replied, there are several ISPs having routing trouble to the particular set of IPs that we have had. We have added another set (but haven’t taken down the old IPs because they still work fine for most folks) which might be better (although YMMV).

Seems pretty clear to me that Kunal Powar is indicating that this could be an issue from AWS side

Ah you might’ve thought that because he used the word “region”. He didn’t mean an AWS region, he meant a geographical region and the ISPs therein.

he was so quick to just conclude that the problem is with my ISP

How do you know he didn’t investigate? Honestly a pretty big issue is your array of assumptions. Please don’t do that, just ask.
Now, to how he concluded that its an ISP issue - we monitor server performance graphs every now and then, he just consulted that. Another hint that this is an ISP issue is that it didn’t affect 100% of the players (if its a server issue, it will affect EVERY PLAYER, not just a specific subset).

I don’t seem to get replies from you guys without making a fuss about it

Please point me to examples. We reply to almost every post, the only exceptions are-

  • folks ask for unbans and cooldown removals
  • folks who DM staff on Discord

We do store demos with voice, but that’ll just tell me that players are complaining about loss, I don’t need that, its not like I don’t trust that you’re facing a problem - what helps is to know WHY you’re facing the issue (which is a WinMTR trace, for example)

Maybe try a different config or some test instance

We do that once we figure out that it is infact an issue with the server (OS, hardware etc). But if we already know that it is an ISP issue, we rather spend time on fixing things that we can do (which is using new IPs and hope it helps). Even there, note how I said “hope it helps”, there’s no guarantee.

Because it is so important, I’ll repeat again, please don’t assume stuff, it doesn’t help anyone :slight_smile:

@anX1 now, two things will help a lot when you experience the loss/choke during high load times (I remember you had mentioned somewhere):

  • Screenshot(s) of netgraph with loss/choke showing, plus points if you can mention the approx hh:mm so we can relate with monitoring
  • WinMTR trace to the server IP at the time you’re facing the issue