I'm trying to figure out if it's just my sheer number of backlog tabs finally doing me in or if it's on Danbooru's side, but for the past while (talking weeks here) Danbooru has been intermittently presenting me with a search timeout on a lot of pages, including just opening a post's page. Today, the site has been aggressively timing out such that I can barely open a single page that has to do with posts, so it's difficult to garden and open posts for viewing/editing without being able to even access a post itself. It takes tens of refreshes to get lucky to even dodge the timeout, and this is with a Builder-level account that has a 9 second timeout window.
Pages that have nothing to do with post entities are completely okay and load within <1s, like the forum.
I haven't noticed anything like this. Edit, I have occasionally timed out when posting an image but very rarely, certainly not as often as what you're describing. (see topic #35810)
Yeah for me it used to be every now and again as well. I have no clue what's different about today but posts are nigh unusable for me. If it has to do with posts, anything from uploading them to editing them to simply opening them times me out, and it takes an absurd number of refreshes and time before I can see anything for just a moment only to get locked in timeout loops again.
but for the past while (talking weeks here) Danbooru has been intermittently presenting me with a search timeout on a lot of pages
For weeks? That's bad. I also got that page time to time but it never was this bad. Could it be a configuration issue on your side? (eg. clear caches, try using another browser, changing DNS settings)
Edit: I jinxed myself everything is taking too long to load now what the hell
Edit 2: I started getting those timeout pages too. Is this a regional issue that is spreading from one to other? I've checked the Background tasks queue and it shows a bunch of queued tasks, could this be related? Reading up on the status it shows I'm connected to the "ina" node, if that helps anything. The status page also shows a pg_dump query that is taking a long time but it should be non-blocking to read from a database?
Edit 3:
it takes an absurd number of refreshes and time before I can see anything for just a moment only to get locked in timeout loops again.
It appears that some of the servers are good while others are bad, I can reach ame ina and gura fine but one server in particular errors out when I try to see it through the status menu, throwing out a type error. Could be the reason why you can do things for a while before getting in a loop.
It's been completely fine on my end, only a little bit of autocomplete slowdown an hour or so ago. Which is weird, I'd expect the database being overloaded would affect everyone.
It's been completely fine on my end, only a little bit of autocomplete slowdown an hour or so ago. Which is weird, I'd expect the database being overloaded would affect everyone.
The specific server 104.23.166.124, gura, gura/danbooru-768bdd4cff-txl6f seems to error time to time and is having a struggle even to load the status page I think that could be something that is affecting the site.
edit: i believe the gura node was being bypassed until recently, once again 104.23.166.125, gura, gura/danbooru-768bdd4cff-r5hzw is timing out or having a response time of around 7-8 seconds.
