For what it’s worth, we see a small spike around 5pm, California time, but it’s usually not the biggest spike of the day, nor very long lasting.
Its happens when waiting for a response feedback related to identification. Normally marking a page as reviewed takes a bit over 10 seconds but during heavy load it might take to around 20s. But then sometimes it takes the 90 seconds for a while before going back to normal some time later (but I usually have a break when that happens). I think there may also be extra delay for annotating (like specifying a sex or flowering) but can’t be certain of that.
Right now it just took 130 seconds to finish reviewing the page (yes, over 2 minutes). Adding a comment took 14 seconds and 15 seconds to add an ID (i.e. about the same time). Its only 2:30PM in California right now. [Perhaps everyone is out with their smart phones uploading observations.]
In my experience there is no need to wait when adding an ID or editing the DQA–just move on and it will save even if you didn’t see it load. As for “mark all reviewed” though, you do have to wait for that one.
I do wait for each ID to show. Sometimes iNat picks the one below what I chose, and the ID is way off!
Odd. That’s never happened to me.
As far as we can tell, there isn’t a regular daily slowdown, but there are a few times a day when things do slow down. We’re looking into making the Identify page a bit less resource hungry.
As far as I can tell, no one’s replied to the both myself and @pisum about whether other parts of the site are also slow during these times.
Today, when I finished identifying a lot of photos, I went to my dashboard. Started going through photos there. After four or five, there was half a photo and nothing else. Website wouldn’t respond. Eventually I went elsewhere. When I returned to iNaturalist, all was well.
When I get the messages about website down or queue full, I’m usually trying to open a taxon page or flag or something.
@tiwane I sort of did say it in my “130 seconds” reply. When the massive delay happens other things are not proportionally slower (i.e. they might only be 50% slower rather than 500-1000% slower).