It's February though, and we've seen at least two outages in the past 3 months, and arguably a lot more generally buggy instances where specific features have been broken (very often without postmortems or any kind of acknowledgement).
I don't think people are surprised to see a website go down; they're surprised to see a website go down often enough that it no longer seems like a particularly noteworthy event. I opened this thread to see what people were saying about the policy change and instead saw people complaining about a service outage, and my response was not, "huh, what on earth is going on at Twitter?" It was, "eh, seems about right."
Twitter always struggled with outages, but they shouldn't be the expected result whenever new features/changes are launched.
Yep, and the event today certainly counts as an outage. I'm being charitable and only counting today's outage and December's (so two), but there's a strong case to be made that I should also be counting things like:
- Spaces going down
- 2FA breaking for a nontrivial number of users and (as far as I know) never getting fixed
It's silly for anyone to claim that Twitter hasn't been less stable over the past few months than it usually is.
----
What makes an "outage" harder to define now is that Twitter basically gives no statements about them and no postmortems. There hasn't been any acknowledgement from Twitter about pretty much anything stability-wise since Musk took over. Which is another way that Twitter today is tangibly different from how it used to operate, but:shrug:
The end result is that "outages" end up being a lot more blurry, and we get conflicting reports from Musk and from other workers about what happened, if we get reports at all. A system breaks and stops working randomly? Twitter's head of safety says it's temporarily down as part of a planned migration. Musk lies and says it never went down at all. Either way, you're not getting a record you can look up later on an uptime page.
But the site does seem to be struggling more often than it used to.
Uh, yes. That's pretty much an outage by definition. Everywhere I have ever worked (and to be clear, that list includes AWS, Dropbox, and Microsoft -- not tiny startups -- amongst others) would consider that to be an outage.
We're squabbling over definitions that change nothing about my main point. If you're arguing that an "outage" means that Twitter needs to go totally down for everyone, then outages are pretty rare. A fair number of the articles above being linked to prove that pre-Musk Twitter had the same problems mostly don't qualify as outages under that definition -- Twitter's 2020 outage didn't result in the entire site going down for everyone. The 2019 outage mostly centered on the US and Europe.
So whatever you want to call that; if you want to say that those events were "disruptions" -- fine, but it is still the case that Twitter has been experiencing "disruptions" at a higher rate than it did in the past, and that those disruptions have involved less communication with users both during and after the events.
Whatever words you want to use for that are fine, I don't really care about that debate. It's still the case that Twitter as a service has been running into technical issues more often than it used to.
We're comparing a world where Twitter being inaccessible for a single hour was viewed as its worst service failure in years, and the current world where people can be blocked from sending Tweets for an entire day and that's not viewed as really all that newsworthy of an event.
> but it is still the case that Twitter has been experiencing "disruptions" at a higher rate than it did in the past, and that those disruptions have involved less communication with users both during and after the events.
Twitter is under scrutiny more now than ever. If so much as an icon doesn’t load first attempt then people can it form or broken.
In the fact before musk bought Twitter. If it didn’t load for 10 minutes in singapore but it did every else no one really cared. Now people are like. OMG ITS DOWN!!!
So no we aren’t squabbling over words. You just can’t redefine outage because you’re upset at Twitter being owned by a twat.
> Twitter is under scrutiny more now than ever. If so much as an icon doesn’t load first attempt then people can it form or broken.
It's difficult for me to point out concrete statistics to refute this because Twitter's current monitoring of disruptions/outages/whatever is fairly abysmal, and Musk has started transparently misrepresenting/lying about them. Unfortunately this means that much of the reporting/monitoring that we're getting is only second hand. But this strikes me as wishful thinking; frankly I don't believe that pre-Musk Twitter had this number of issues.
Is it possible that there's increased scrutiny, so more people are reporting problems on Down Detector? I guess, it's not impossible. But again, this gets back to what I was mentioning above -- we shouldn't have to rely on Down Detector for numbers, but we do because Twitter has basically completely abandoned self-reporting its own stats.
I can't completely disprove or dismiss the theory that users are just reporting issues more often now, but I don't think it's the most likely explanation available, and I don't see a ton of evidence to back it up -- the simpler explanation is that people are reporting more issues because there are more issues.
----
Secondly, I do think we are squabbling over words. It is still the case that if you define "outage" as a worldwide global site unavailability, the majority of the previous links listed above as evidence that this is nothing new don't quality as outages.
We've shifted from saying, "Twitter had outages in the past, this is nothing new" to "okay, the stuff happening now isn't an outage (but somehow the past events still count)", to "actually, things were much worse in the past than you remember and you just never noticed."
But... again, I just don't buy it. Look, 10 minutes of downtime is not the same thing as the private API going down across a substantial portion of the US for an entire day (at least a day, we still don't know for sure if the problem is fully fixed because, again, communication is nonexistent).
Twitter went down for an hour pre-Musk and the media reported it as the worst outage the site had suffered since 2016. Like... come on, it's obvious the site is having issues more frequently than it used to. My opinion of Musk is irrelevant to that observation, the guy could be a saint, and it's still obvious that the site is less stable.
I would argue more than two. 2FA was broken globally for a bit, AIUI, and in some geographies is still broken. That’s effectively an outage for 2fa users. There’s also a lot of general flakiness which you wouldn’t call and outage, but are very much a return to the Failwhale Era.
That's a wild exaggeration. I would experience occasional hiccups once or twice a year. I'm usually pretty active, tweet or reply a few hundred times a month.
reply