Ops Bug: Why does the site keep disappearing?
-
Site has been up and down a bit this evening. Are y'all applying updates or is there some reliability problem?
And as my ex girlfriends would say "it's not me, it's you" - I do, of course, try other sites when this one goes awol, and they work.
Outage seems to be for ~a couple of minutes at a time.
-
Very weird, updates are meant to be seamless.
Anyone else experiencing this?
-
I just updated about 3 minutes ago.
-
No repro here, but I've not been on the site 60x60...
-
I was seeing quite a lot of problems resolving the DNS entry sometime this week, but I don't know which day I was seeing it on. It's been that sort of week.
-
It did go down for me last night (PDT)
-
last night (PDT)
That time format is almost as useful as the one in the upper right of the post! I think it's the timezone that really sells that it was night.
Feature request: label post timestamps with more descriptive language, such as "way back when" and "in the dark times".
-
That time format is almost as useful as the one in the upper right of the post! I think it's the timezone that really sells that it was night.
Feature request: label post timestamps with more descriptive language, such as "way back when" and "in the dark times".
You must be a friggin' blast at parties.
-
I just updated about 3 minutes ago.
That would fit with what I saw. And was it updated once more, earlier in the day?
I'm not even a gold-card-holding frequent flyer, connecting maybe four times a day for 15 minutes at a time, yet I've noticed this a handful of times both from work and at home.
Symptoms are; page half draws then spins. F5 = 404 (eventually), new tab = 404, try again in 2 minutes and it's back.
-
You must be a friggin' blast at parties.
You must be extremely sensitive.
If that mild joke partially at your expense rubbed you the wrong way, this may not be the best forum for you?
Filed under: The smiley cat implies a post made in a jovial manner
-
Been trying to login to these forums all day from work, Discourse just gives me the spinner on the login dialog and nothing happens - I left it for 15 minutes or something stupid and it still hadn't logged in so I gave up. Related?
-
-
This site's been down today more time than it's been up.
-
I haven't noticed it being down at all.
-
-
That is really odd. Outside of the occasional version upgrayyeds, I can't explain any outages, and this site is hosted on Digital Ocean. Any way we can diagnose these issues @sam?
-
I thought that the main site and the "what" subdomain were hosted on different servers. The main site being unavailable can't be anything to do with discourse can it?
-
This is a worse barrier to reading than infinite scrolling.
-
I agree @error -- this site is not hosted on our dedicated servers at he.net -- it's hosted on Digital Ocean 2GB / 2 cpu tier.
I added what.thedailywtf.com to our Pingdom monitoring with a 1 minute interval so I can get a sense of what's going on.
-
I added what.thedailywtf.com to our Pingdom monitoring with a 1 minute interval so I can get a sense of what's going on.
Sounds like a plan. We had an issue where a customer alleged that their app "kept going down" but we suspected they were just looking to get out of the contract.
We configured a harmless walk-the-stack method on a non-publicised URL that exercised, well, the whole stack, then set pingdom up against it.
We were able to prove that the allegations of the admin were unfounded and relationships gradually improved.As the OP, I'm not suggesting you set out to prove me wrong (!) but it may be worth you setting up a similar technical spike on an unpublished URL, with logging-level=fucking-everything, and throw pingdom at it?
-
When I was having problems a few days ago, it was at the host resolution (i.e., DNS lookup) stage. Like there were some problems with the provider of that service for thedailywtf.com, or possibly for the registrar. I don't do much fiddling of DNS so I can't diagnose, but DNS trouble caused us to move to a completely different set of providers and that helped a lot.
-
I recently deployed here (tab counts on the front page are not insane) and was furiously hitting the refresh button during the deploy. Was unable to break anything.
I do wonder if this is DNS related.
-
Seems plausible to me. I just wanted to make sure our recommended Digital Ocean docker config didn't have any weirdnesses.
-
obviously Comcast and Verizon trying to get TDWTF to purchase a "fast lane"
-
-
When I was having problems a few days ago, it was at the host resolution (i.e., DNS lookup) stage.
We had one client suddenly having trouble accessing the web app hosted with us, and they were using OpenDNS. I don't remember how it got resolved.
-
We had one client suddenly having trouble accessing the web app hosted with us, and they were using OpenDNS. I don't remember how it got resolved.
By not using OpenDNS perhaps?....
-
Could be, or OpenDNs updated itself or we made a tiny change in our domain names. I honestly can't remember.
-
Could be, or OpenDNs updated itself or we made a tiny change in our domain names. I honestly can't remember.
-
this site is not hosted on our dedicated servers at he.net
The dedicated server is with Hivelocity, not he.net.
it's hosted on Digital Ocean
They're not the most reliable host, I saw lots of random short network outages when I was testing out their service for my own hosting. I ended up using BuyVM since their servers didn't feel as oversold (more responsive and better disk IO throughput).
-
http://cachecheck.opendns.com/
Maybe I should have also added a giant
#*SHRUG* i dont care
to my post i dunno