Congratulations on your glorious four hours of downtime
-
I'd like to thank everyone here who is in the pocket of Big Discourse. You know who you are.
-
Big Discourse sure are proud of all their ⑨s...
-
Allow me to reply with a poem
One night I dreamed I was walking along the beach with Discourse.
Many scenes from my life flashed across the sky.
In each scene I noticed footprints in the sand.
Sometimes there were two sets of footprints,
other times there were one set of footprints.This bothered me because I noticed
that during the low periods of my life,
when I was suffering from
anguish, sorrow or defeat,
I could see only one set of footprints.So I said to the Discourse,
"You promised me Discourse,
that if I used you,
you would always walk beside me.
But I have noticed that during
the most trying periods of my life
there have only been one
set of footprints in the sand.
Why, when I needed you most,
you have not been there for me?"The Discourse replied,
"The times when you have
seen only one set of footprints,
is when I carried you."
-
"Amen"
-
That outages number is a tad disappointing aswell...
Over what period does that cover?
-
So TDWTF is in danger of bringing down the averages from 3 nines to 2?
I'm trying to understand.
-
the report covers November.
25 outages with a total of 4 hours down time.
average 9.6 minutes downtime per outage.
total outages on all sites monitored in the report: 38.
% outages at TDWTF : 65.blahblahblah% (or roughly 2 out of every three outages)
-
The average response times are absolute rubbish.
-
It's a small price to pay. Once the page is loaded, you can see Infinite Posts!™ No need to click next page every time.
Of course it still takes a couple of seconds to load the next chunk. And it happens much more often than you had to click next page in the previous forum, meaning you still lose more time. But you save one click every 50 posts!
-
@Intercourse said:
The average response times are absolute rubbish.
It's an outrage!Filed under: I just wanted to use that word.
-
Pingdom needs an Android emulator for their real user sims, then we could see some buffer overflow on the response times. (Yes, I know it would show outage before that)
-
So, why do we have square the amounts of outages than the next big competitor (try.discourse)? And don't you guys need to restart try.discourse.org every once in a while for all the updates you post on that?
Also, we had (/ still have sometimes?) the white-post problem. As a forum that should count towards outage.
Also that Downtime number seems fake? What with being a perfect 4 hours?Filed Under: We have the worst uptime.... we need bigger pockets!
-
Oh hey, what a shocking revelation: Discourse is a piece of shit. Who saw that coming.
-
I think it might be related to some sort of gross misconfiguration of the logging coupled to the way some people use bots. At least some of the times that DC has been down, it's been because we've filled up the disk (presumably with logging info) and have needed emergency Staff attention.
Plus the white post phenomenon, which was pure Rubyfail.
-
Yeah there have been a few things like that:
-
better default nginx template rate limiting and docker image log rolling due to excess bots, caused disk to fill with logs
-
disk filling up due to extra backups
-
posts failing to render (current theory is nokogiri freakout, v8 has been ruled out)
I am not sure how Pingdom calculates response times but I think they average from the whole world and all their ping sites, of which they have a lot. I remember the ping response times being pretty bad for Stack Overflow back in the day as well and that is a very, very fast site. Thanks compiled .net code!
-
-
Thanks compiled .net code!
One of the biggest WTFs around here is how you're aware of a programming environment that doesn't suck ass and can be used to create decent websites, and somehow built this in a giant open source shitpile instead.
-
Meh, I think it was the right tool for the job in this case. For projects like this you want to be able to iterate quickly (although, you should test a bit more thoroughly before releasing) and there is a huge community around RoR. Ruby support is also pretty ubiquitous these days, the hosting is cheap and with the advent of Docker, anyone can have it up and running in a short period of time. The same cannot be said of .Net code, the price of entry would be a lot more expensive and it would be a lot harder for a novice to get up and running, not to mention the hosting would be a lot more expensive.
The only thing more ubiquitous these days for the purposes that they wanted to address would be PHP. So which would you rather have?
-
Is that a big discourse in your pocket, or are you just cocky?
-
Is that a big discourse in your pocket, or are you just cocky?
I assume that comment was directed at me? Nah, not in the pocket of Big Discourse. Even a broken clock is right twice a day, they can't do everything wrong. ;)
-
@Intercourse said:
For projects like this you want to be able to iterate quickly
And that has what to do with choice of language?
@Intercourse said:
and there is a huge community around RoR.
Then maybe King Asshole there should see if he can find some good developers to work on this shitpile.
Of course, what he really needs, since his moron devs are obviously incompetent at it, is a few fucking testers.
@Intercourse said:
the hosting is cheap and with the advent of Docker, anyone can have it up and running in a short period of time. The same cannot be said of .Net code, the price of entry would be a lot more expensive and it would be a lot harder for a novice to get up and running, not to mention the hosting would be a lot more expensive.
Any hosting that allows Docker allows Mono. In fact I'm pretty sure you can just run Mono inside your Docker VM or whatever they call it.
@Intercourse said:
So which would you rather have?
A forum that fucking works? That has less than 34 new bugs a week?
-
And that has what to do with choice of language?
RoR makes it a lot faster to deploy new features, etc. At least for me and those I know. If you take two projects and develop one in RoR and another in .Net, RoR will usually win for speed of development. To each their own though.
Then maybe King Asshole there should see if he can find some good developers to work on this shitpile.
Sam seems competent. Not too familiar with the others. And really, I guess it would be impossible for me to really gauge their competence without working with them.
what he really needs...is a few fucking testers.
I would certainly agree with that and I feel like most issues that crop up should have tests written for them to verify that they are fixed. Then, if you ever have a regression, it should be caught by automated testing. TDD only works if you follow the procedures.
Any hosting that allows Docker allows Mono. In fact I'm pretty sure you can just run Mono inside your Docker VM or whatever they call it.
Yeah, you are correct there. I always forget about Mono. When I think .Net, I think Windows Server.
A forum that fucking works? That has less than 34 new bugs a week?
Agreed. And far too many of those are regressions which should have been caught by testing. To be fair though, I believe we are not on the stable branch. Also, I speculate that we are likely on the bleeding edge so that we can catch those bugs as some sort of arrangement between Jeff and Alex. That is a WAG though.
-
@Intercourse said:
I assume that comment was directed at me?
What, no, sorry, I just thought it was a funny thing to say. If anything I guess it was inspired by that ‘footprints’ nonsense upthread.
-
What, no, sorry, I just thought it was a funny thing to say. If anything I guess it was inspired by that ‘footprints’ nonsense upthread.
Ahhhhh, makes sense now. Not that I was offended or anything.
Always remember, not quoting every...single...thing, is Doing It Wrong. ;)
-
@Intercourse said:
RoR makes it a lot faster to deploy new features, etc. At least for me and those I know.
You'd have to cite a hell of a study between, say, Rails and MVC. A hell of a study. Because that's utter gibberish. Plus .net environments have those nice things like "working debuggers" that make writing code so much easier.
@Intercourse said:
If you take two projects and develop one in RoR and another in .Net, RoR will usually win for speed of development.
Proven with... what evidence?
@Intercourse said:
Yeah, you are correct there. I always forget about Mono. When I think .Net, I think Windows Server.
Well, I'm sorry you're a fucking moron then. I guess.
I really appreciate speed-of-development advice from a moron who doesn't even know Mono exists, BTW. That's evidence of a great mind with well-rounded IT knowledge to draw on.
-
The math is bad on that. 4 hours of downtime during Nov is not 99.45% uptime. Nov has 30 days, which is 720 hours. With 4 hours of downtime, that's 716 hours of uptime. 716 / 720 = 0.99444444 or 99.44%. Looks like pingdom needs some math help.
-
maybe they calculated that answer on a Pentium FPU
-
@Intercourse said:
For projects like this you want to be able to iterate quickly
You can do that in .NET. I've got a .NET project where we are sometimes updating production 2 or three times a week. With a team of two.
@Intercourse said:
The same cannot be said of .Net code, the price of entry would be a lot more expensive
For devs? Not true. You can use the appropriate express edition to develop commercial apps for free.
@Intercourse said:
it would be a lot harder for a novice to get up and running, not to mention the hosting would be a lot more expensive.
First off, CDCK offers hosting services for the novices you mention, so that's not a huge issue. As for hosting costs, you can find Windows hosting starting at $10/month, and there's always mono if you want to try hosting in another environment.
-
Does this uptime count the times when we get stuff like magical empty posts? I just had one. Even rebaking didn't do anything. I had to make an edit to make it visible:
Fuck...and the next one, too:
....and then:
So....are we up or not right now?
EDIT: GODDAMNIT, THIS ONE WAS FUCKING INVISIBLE TOO!
-
Your'e getting notifications, but you can't read this!
-
Does this uptime count the times when we get stuff like magical empty posts? I just had one. Even rebaking didn't do anything. I had to make an edit to make it visible:
Fuck...and the next one, too:
....and then:
So....are we up or not right now?
EDIT: GODDAMNIT, THIS ONE WAS FUCKING INVISIBLE TOO!
@sam @codinghorror
Your'e getting notifications, but you can't read this!
hmm... last time my quote was visible (at least to me) does it work this time too?
EDIT: Yes, until i refreshed the page.....
-
Oh hey this again. Downtime that doesn't count as downtime? Let's maybe change the title here to "glorious two MINUTES of uptime" because that seems more accurate.
-
-
@accalia - Days Since Last Discourse Bug: null
-
@accalia - Days Since Last Discourse Bug: <i>null</i>
<!-- Posted by SockBot 0.13.0 "Devious Daine" on Sun Dec 07 2014 03:33:59 GMT+0000 (UTC)-->
Oh sure.... YOUR post cooks correctly.....
or so it appears to me....
http://what.thedailywtf.com/t/the-amazon-likes-dispensary/1000/30946
http://what.thedailywtf.com/t/the-amazon-likes-dispensary/1000/30947
http://what.thedailywtf.com/t/the-amazon-likes-dispensary/1000/30948
http://what.thedailywtf.com/t/the-amazon-likes-dispensary/1000/30949looks like some of the bots aren't affected byt the white post issue....
why?
-
looks like some of the bots aren't affected byt the white post issue....
why?
Discoursistency!
-
And yet when I quoted blakeyrant in another thread, just like the post above this one, it's invisible too!
-
Shit I was on my phone I didn't see this post when I did it :(
-
You'd have to cite a hell of a study between, say, Rails and MVC. A hell of a study. Because that's utter gibberish. Plus .net environments have those nice things like "working debuggers" that make writing code so much easier.
Meh, I can call this one a draw. I can only speak from my own experience and I CBA to look up a real study. You are free to though, and if I am wrong (unlike you) I will admit I am wrong.
Proven with... what evidence?
My own anecdotal evidence, which my my own standards is admittedly not-admissible. Draw.
Well, I'm sorry you're a fucking moron then. I guess.
I really appreciate speed-of-development advice from a moron who doesn't even know Mono exists, BTW. That's evidence of a great mind with well-rounded IT knowledge to draw on.
You fucking cunt, did I say that I was unaware of Mono? Yes, saying that:
@Intercourse said:
I always forget about Mono.
Is the same as saying, "I am completely unaware of Mono, thank you for enlightening me." You need to have your shoulder aliens checked out. Those motherfuckers are apparently schizophrenic. You fucking cunt.
You can do that in .NET. I've got a .NET project where we are sometimes updating production 2 or three times a week. With a team of two.
Fair enough, I was speaking from my own experience. Right now we are in the alpha stages of a RoR project that I started. We are still early enough on that dev = production. Right now when an alpha-user requests a feature, we implement it, pull from Git and then it is live. Updating production happens A LOT, per week. If it passes tests, it goes to production to be evaluated. This is very early on though.
For devs? Not true. You can use the appropriate express edition to develop commercial apps for free.
Meh, I was talking about the amount of time it takes to get to profitability. As far as dev tools go, I could really give a shit less. Those are one time costs. In my frame of reference, I am talking about the amount of time (and time=money) that it takes to get to a billable consumer. I get your point though.
First off, CDCK offers hosting services for the novices you mention, so that's not a huge issue. As for hosting costs, you can find Windows hosting starting at $10/month, and there's always mono if you want to try hosting in another environment.
Fair enough, but it depends upon how advanced your application is. Off the top of my head, as a guesstimate, I would think that a .Net (Windows Server) solution would require a VPS, which will cost you more. A Digital Ocean Discourse instance will cost you the same amount of money though. I call it a draw.
@blakeyrat, you should notice the difference in tone. When I talk to @abarker, I am very cordial. When I talk to you, I am a bit of a cock, because you are a cunt when you reply to people. You don't know everything, and you need to have those shoulder aliens checked out. You seem to be wrong a lot lately.
-
How is it possible that this bug is not fixed?
-
Oh, fuck you @codinghorror, I was taking up for you and then your creation shit itself like this. Fuck it, I give up.
-
Nothing in the logs here, this error is killing me
-
Experimenting, just nuked sidekiq, which auto spawned, and left parent process as it
-
Interesting ... so we can now post correctly, so it appears the background sidekiq job is responsible for the blanking
-
Added this patch, I am pretty sure this is where posts are being blanked
Will watch logs. (note, not deployed here yet)
-
-
-
For devs? Not true. You can use the appropriate express edition to develop commercial apps for free.
There is also the VS2013 community edition which is just one of the Professional versions rebadged.
More here on a lot of the interop stuff they are working on:
-
@Intercourse said:
You seem to be wrong a lot lately.
That's a funny thing to put at the end of a post where you admitted I was right about 3 times.
-
That joke is 20 years old now.
-
@Intercourse said:
Fair enough, I was speaking from my own experience. Right now we are in the alpha stages of a RoR project that I started. We are still early enough on that dev = production. Right now when an alpha-user requests a feature, we implement it, pull from Git and then it is live. Updating production happens A LOT, per week. If it passes tests, it goes to production to be evaluated. This is very early on though.
Once again... what does this have to do with language? You could set up your favorite source control system to trigger deployments to any environment you want on every commit if you wanted to. I can't think of any language that wouldn't allow this. We have most of our .Net projects set up to do this for our test environments.