same the site very often says perfomance issues cant search or performance issues so youre seeing it as if you werent logged in id love to donate because this is the only place that has new scripts coming out constantly.
There are donation links at the bottom of each page but it is limited to crypto unfortunately.
That should be the next challenge to tackle, IMO.
Whether it’s a Patreon (“@hugecat is developing an Art Community Portal”) where you just post (basically) Patch Notes and Site Traffic Notes about the site, without ever actually naming or linking the site, or a SubscribeStar, or even a freakin’ OnlyFans where you post Patch Notes and Site Traffic Notes - that gives you an avenue to accept non-crypto cash, and no matter which one you choose, and no matter how much cut the site takes, the number of people who can sign up for a $5/mo RECURRING subscription vastly outweighs the number of people who can send Crypto, ever.
Supporting Data: there are 2998 people who have visited the site more than 3 times this week (since October 8)
If only 1% of them signed up for a $5/mo Patreon/SubscribeStar/OnlyFans, and after the site cut and whatever losses conversion out of USD brings (let’s say it’s 50% drag overall) That’s still $75/mo - which, if I recall correctly, is what you said it costs you out of pocket to host the site a while back.
2 Likes
Fixt. 
doubt it, likely just due to higher traffic?
if that’s a common issue though, let me know. I haven’t heard too many complaints about that in a while.
It has been good until maybe 1-2 weeks ago when it turned bad really fast. When writing this a page takes 20-30 seconds to load, but no message about heavy load is shown as it has other nights when the server has been a much more responsive compared to today.
Edit: and after writing this message the server is quite responsive again. strange.
@hugecat do you have access to the site’s performance monitoring tools? You should be able to see if there is a trend of high CPU, disk, RAM, or network loads.
I do see CPU spikes, but imo the graphs are not super helpful. CPU usually hovers high and ram usage is always consistently like 90%+.
I can restart the server maybe tomorrow, see if that fixes anything. Otherwise I’ll take a look at upgrading the server again.
Got a couple of 504 gateway timeout yesterday if that is of any help.
same. its been so terrible lately.
Was the server rebooted already? Response has been closer to normal for the last few hours.
Site performance is good / normal in the morning / noon but gets worse late afternoon / in the evening (European time zone). That’s for the last 2 weeks or so.
Evening EU-time is the worst. Given that the server has been running very well for many months it is hard to believe that increased traffic should be the culprit. Traffic don’t tend to change that fast unless the site has been promoted somewhere and been given a lot of attention lately. Anyways, aren’t traffic statistics available by the web host?
BTW, as someone already mentioned the “site is under extreme load” message also started to pop up recently, even when the site was performing well. There is a thread about that with some tips on how to configure workers and some db settings. The one asking got this message despite having resources available on the host. This was solved with a small reconfiguration of discourse. In the link there is another link with more in-depth info on how to configure worker threads
There also seems that others experience slow down after the upgrade to the latest version. The thread doesn’t give any answers, but it mentions a built-in mini-profiler that might be useful.
1 Like
restarted and updated discourse too, let me know how things go
Feels sluggish around 7pm EU time. Took unusually long time to login for example and pages takes up to 7 seconds, but some loads almost instantly so I guess it is to early to say. Did you check that the server use an appropriate number of workers to serve the discourse app too after the upgrade?
yes, it’s following the recommended % for db shared buffer and number of unicorn workers.
A long time ago I tried to investigate if there was some rogue process or something that was tanking performance, but I couldn’t find anything suspicious and concluded that we just needed to upgrade machines. I’d guess it’s the same situation right now.
Just out of curiosity, is ES hosted on a dedicated server or is it on a vm?
currently on digitalocean droplets, basic droplets, premium amd cpu