NINETY DAYS

NINETY INCIDENTS

NINETY PERCENT

YOU PAID FOR ALL FIVE NINES BUT YOU’LL ONLY NEED THE EDGE

#github

@0xabad1dea There are five nines on the screenshot, I don't see any problem.
@dos @0xabad1dea might have been 6 if the waited two days!
@0xabad1dea is that a sequel to "Nine Hours, Nine Persons, Nine Doors"?
@ratsnakegames @0xabad1dea nine hundred ninety nine thousand, nine hundred seconds. how do you measure... yearly uptime?
now i'm thinking of the cover of the DS game "999: nine hours, nine persons, nine doors" but it's "909090: ninety days, ninety incidents, ninety percent"
@0xabad1dea i bet i know what's going on here
@erikarn @0xabad1dea synthetic text extrusion doesn't actually make great software.
@0xabad1dea Before you can improve your process, you have to make it consistent.
@0xabad1dea Give it time and they'll get it to nine fives.
@0xabad1dea This is what happens when the “official” Whitehouse grift app loads JavaScript from some random GitHub pages.
Shut the whole site down for a month to fix stuff and you'd still have a better uptime
@0xabad1dea Maybe they meant 9 5s?
@0xabad1dea I'm sure #AiSlop can't be the problem! /s
@0xabad1dea www yeah baybeeen looks at dem nines!
@0xabad1dea You're looking at this all wrong. If you express it as eighty nine point nine repeating, it's infinite nines. Microslop wins this one!
@0xabad1dea fun fact: that's about a month and a week's worth of downtime.
@0xabad1dea does 89% uptime count as one nine ?

@gkrnours @0xabad1dea Ooh ooh, I can answer this!

Given that 100% is hard to reach and 0% is not, we can use a simpler version of the logit function and take log₁₀(1÷(1−.89)) = ~0.96 which means it's still a tad below 1 "nine".

(The full logit function is log₁₀(x÷(1−x)) and is useful when both 0% and 100% are hard/impossible to reach.)

@0xabad1dea

hey, 89.99% reliability is still 3 9s reliable :3
@0xabad1dea
I paid for the whole uptime status bar, so I am going to use the whole uptime status bar!
@0xabad1dea I'm more of a nine fives kind of sysop
@0xabad1dea this rocks, pick any random homelab on fedi and it's probably destroying github lol

@0xabad1dea Once upon a time, at a small entertainment company that rhymes with 'misery', I was responsible for the operations of the Github Product.

It came as a vmdk. No documentation.

I was told "Backup" was as simple as snapshotting the VM.

It ran gunicorn.

There was no tuning, monitoring or built-in telemetry.

They refused to give us an extension on licensing when they forgot to cash our check, and shut it off from remote.

Every call was with techbros.

I cannot stress enough, how happy this makes me.

@0xabad1dea THERE ARE FOUR NINES!!!
@0xabad1dea Grok, is this true? 🤖💩
@0xabad1dea insanity. when i worked for microsoft they would have fired me and then launched whatever was left into the sun if i let this go on for four hours much less ninety goddamn days
@0xabad1dea I nearly choked on my cheese. Good job!
@0xabad1dea Where did this come from? status.github.com paints a very different picture, though also not particularly great honestly…
@slembcke @0xabad1dea This is from the "missing GH status page": https://mrshu.github.io/github-statuses/
The Missing GitHub Status Page

Historical GitHub uptime reconstructed from archived status data.

@slembcke to confirm what someone else said while I was asleep, it’s an independent third party tracker called the “missing” github status page. https://mrshu.github.io/github-statuses/

they track ten subsystems separately, which are each in the 96% to 99.9% range, then overlay all those incidents on the same timeline to arrive at 90% overall.

I do not have any particular stance on whether github’s own tracker or this third party one is more fair and accurate, beyond “90-90-90 is hilarious” and observing only one of them has a contractual stake in possibly sometimes downplaying issues a little bit.

The Missing GitHub Status Page

Historical GitHub uptime reconstructed from archived status data.

@0xabad1dea @slembcke if each subsystem is crucial for github to work, you multiply those uptimes and get 90.
If the subsystems are independent then can't do that.
If the CI system is down, you can't say github is down.
I would then merge the timelines and get some metric "how long has there been no issue at all".
That can be very low and the website is still fine.

@0xabad1dea There's a 996 joke in here somewhere, I swear.

Edit: 996: one 9 of uptime, 9-0 incidents, and the third nine in "90 days" is upside down lol. We can workshop it.

@0xabad1dea they're right there, idk what your problem is
@emily @0xabad1dea I don't know if I should make a Plan 9 or a Touhou reference, so https://analognowhere.com/_/lxhtmg/
this fall cannot front

anime girls

@0xabad1dea oh, no! Just a few more days, and it would also have been the 90th day of the year (the 31st, if I counted correctly)

@0xabad1dea @adrianco

Somebody just has to type "continue" into the right co-pilot chat at Github.

@0xabad1dea pretty sure this means https://solar.lowtechmagazine.com/ now has a higher uptime percentage, despite running on an SBC on a solar panel on a balcony
LOW←TECH MAGAZINE

This is a solar-powered website, which means it sometimes goes offline

LOW←TECH MAGAZINE
@0xabad1dea

We see GitHub is still working on that second nine of uptime:
https://transfem.social/notes/ak8mfds9u2eg00ac
Site Reliability Enby (@SiteRelEnby)

We see github is still working on that second nine of uptime. (📎1)

TransFem Space
@0xabad1dea We're just sloply moving to a new industry standard of nine fives.
@0xabad1dea I don't know, but hot take... 1 is not 5... Though what would I know... I'm not good at math. Apparently. In other news... 3 is not 4... These are interesting things that seem to be hard for others to know, or maybe I'm just the stupid person who doesn't understand math.
@0xabad1dea That's worse uptime than most of the services I run for myself. :P
@0xabad1dea this new github coopilot feature will improve microslop copilothub copilot impersonating ability, now you will be able to hand out your private projects to microslop copilot without copilothub breaking.
@0xabad1dea It's my perpetual joke about Monkeybrains and their eight eights of uptime but the load-bearing structure of the interweb.

@0xabad1dea the green bits are 100%

It’s 99.999% but we’re not telling you the timeframe.

@SecurityWriter @0xabad1dea 90% of the time it works 100% of the time

@SecurityWriter @0xabad1dea

Or they select very carefully upon the calculation:
it's a difference if you count ALL services working and ANY disruption, even that of a single service/server as fail, or whether you count a single server failure against the total number.

If you have 100 (virtual) servers you can keep one server swicthed off ALL THE TIME (0% individual availability) and still have 99% availability overall.

@vampirdaddy @SecurityWriter @0xabad1dea I once worked at a company I liked that said "anyone promising 100% uptime is lying"

They got bought out and ruined by a company that said "100% UPTIME*" on the wall

and in tiny tiny writing "*core network uptime" way down in small print in the basement

@vampirdaddy @SecurityWriter @0xabad1dea Because, technically, they had a very large core network and they'd had a COMPLETE nationwide network outage.

Of course this bullshit did not fly anytime they had a LOCAL outage.

@0xabad1dea the yolocolo at european chaos/hacker events has used "9.9999% is also five nines" in some posters...

maybe that's the logic Github's operating by too

@0xabad1dea somebody should organize a class action
@0xabad1dea you can get more 9s of uptime is the first number is an 8