Indeed. Just happened this Monday (morning) for me.
Yep. at 5:52 GMT got the Code 500 error, which I’ve seen numerous times in the past at that time day, mostly. Didn’t retry until just now.
This seems to be happening almost every day at this point. Just got it again this morning.
1 Like
Seems like a weekend traffic jam to me.
You notice it at weekends because I sleep more and am not around to reboot the server as soon as it goes offline 
11 Likes
This is something we’re trying to fix, but trying to co-ordinate 3-4 people with a number of responsibilities to fix it is surprisingly hard when this is going to take a while to sort 
7 Likes
Rosemary, thank you for your work on this. We appreciate it.
4 Likes
Oh, that explains a lot! 
I’m so glad you’re on the case! Have you identified the issue yet?
Sleep is good. Fun random story - I actually was working with a server back quite a ways that had a really, really bizarre memory leak. We actually scheduled a cron job to reboot it every couple days in the wee hours of the morning. 
1 Like
@RosemaryOrchard Maybe you already know, but we could set up an automated restart every N hours or Nth hour of the day via crontab (or something else) until we find the root cause/fix for this. That way, you can sleep more every day and MPU users can continue to geek out on the forums 
I do know that, but that’s a patch rather than a fix for the actual problem. We need to migrate the forum to some newer shinier servers which will allow us to run some other much needed migrations in the back end, but it’s not something we can just do at the drop of a hat 
7 Likes
The forum is being moved to a different hosting provider?
Probably a G5 tower in my studio
14 Likes
(Fine) Grated cheese for everyone! I’m in! 
Yeah, upgrade to a better hosting plan! 
Almost the entire forums existence I’ve noticed quite frequently times when it is extremely slow to load or down. No real judgement just always wondered why it is.
I do! It often automatically reboots itself, but as someone who does this for a living I also immensely dislike throwing a patch out to fix something without really knowing what the problem is. Every time the forums go down I do read through the log files trying to identify the cause, but I’m neither a Discourse, Docker, nor Ruby expert which sadly makes this more difficult.
6 Likes
Time to host it on Cloud (AWS, Azure) with Load Balancing and Auto Scaling enabled.
I can understand the situation, being myself in Software. But data to client is data. Whether it is junk or not, when promised a service and is paid, it should be honored. Junk for you may not be junk for him. May be that’s the only time in his time zone the person is able to clean up the junk. Or may be the person just wants to get his money worth by making un realistic demands to what meagre amount he paid.
At the same time what if a $2500 client asks for the same junk email
Something that really helped me with a similar problem was enabling the slow query log in MySQL. Knowing which queries were causing issues was huge, especially when I discovered a weird SQL injection bug that was allowing people to throw “sleep” commands at the database.