So last night's fight broke BoxingScene

[.!WAR MIKEY!.]

Is this site ever going to upgrade so that it can handle the traffic of a big boxing event?

The times when it could be attracting new posters and new eyeballs for ad revenue?

This site sucks on big fight nights, It become Boxingscene.geocities.com lmfao.

[juggernaut666]

What kind of statement is this ? If anything THEY SHOULD BE ABLE TO HANDLE THE INCREASED TRAFFIC POST FIGHT, unless of course they don't care about the lost ad revenue because people cant log on.

Then theyre losers . You have winners and losers ………………...


Honestly I have no idea what hes really talking about I just cant stand him !

[O Prophet]

It happens after every big fight

[Mooshashi]

Nothing wrong with the Scenes' servers...look fine to me.

[Combat Talk Radio]

Nice. You a Cloud / infrastructure Engineer ?

You thinking have an ngnix front end and hosting on AWS ?

General tech consultant.

First step is architecture review. If the site is running on some dude's Dell at home, we got a lot of work to do. But it feels like that.

It's funny too, you got Cloudflare but not using its function correctly. A single server behind Cloudflare is not only a waste of money, it's defeating the purpose of the tool (uptime assurance).

I actually think someone restricts traffic on purpose. Basically has a DoS set up to prevent or throttle connects when the server is near some metered amount of traffic. But there are better ways to do that instead of just making it not respond.

[j0zef]

This happens after every big event. Hell, even during not so big events it stops working/slows down to a crawl sometimes.

[BIGPOPPAPUMP]

Actually the site never went down during the Canelo-GGG events or Mayweather-McGregor.

We were monitoring the traffic, but there was a massive surge of traffic around the time of the Martin-Washington fight, which overloaded the server.

We are on cloud servers, so there is no need to upgrade when additional resources are added in via cloud. More resources were added to the site and the site stabilized at the end of Wilder-Fury and from that point had no issues.

Having said that, we are moving to Amazon servers in the coming months.

[BIGPOPPAPUMP]

General tech consultant.

First step is architecture review. If the site is running on some dude's Dell at home, we got a lot of work to do. But it feels like that.

It's funny too, you got Cloudflare but not using its function correctly. A single server behind Cloudflare is not only a waste of money, it's defeating the purpose of the tool (uptime assurance).

I actually think someone restricts traffic on purpose. Basically has a DoS set up to prevent or throttle connects when the server is near some metered amount of traffic. But there are better ways to do that instead of just making it not respond.

The site runs on numerous servers, all on a cloud structure. Two web, two mysql, load balancer, caching server, slave server. Six servers in all. The site has not run on a single server in at least a decade, maybe longer.

[i_am_a_champ]

General tech consultant.

First step is architecture review. If the site is running on some dude's Dell at home, we got a lot of work to do. But it feels like that.

It's funny too, you got Cloudflare but not using its function correctly. A single server behind Cloudflare is not only a waste of money, it's defeating the purpose of the tool (uptime assurance).

I actually think someone restricts traffic on purpose. Basically has a DoS set up to prevent or throttle connects when the server is near some metered amount of traffic. But there are better ways to do that instead of just making it not respond.

Nicr man sounds like you know your stuff . i did notice some 500 status codes , 503 unavail and 504 gate way timeouts probably due to throttled bandwidth like u mentionef

Is there are article or tutorial u xan point me that sums up all u described? One u recommend

[Combat Talk Radio]

The site runs on numerous servers, all on a cloud structure. Two web, two mysql, load balancer, caching server, slave server. Six servers in all. The site has not run on a single server in at least a decade, maybe longer.

Sounds like a decent footprint...for about 10k users.

It makes me wonder if your load balancer simply isn't optimized for traffic routing according to worst-case (basic round robin is only effective if you've got more than two servers doing it).

I'm assuming you have the caching server configured to serve up static information in the event the two nodes don't respond - otherwise why have a caching server, you know? And if that's the config, the site should still respond, even if the forums display a static "Forums are temporarily offline" message as 269 or 503. Usually points to a timeout misconfiguration (if you can't respond to the request in 15 seconds or less, return 503 as ____.htm, for example).

IN my experience, the site doesn't even respond with a code. It just sits spinning. I haven't seen that symptom since I was working for the @Home Network back in the 90's.

Either way it sounds like some backend changes are coming, so as long as AJ/Fury or a Mayweather/Canelo rematch doesn't get booked we should be fine. Heh