Meta outage

Mailman List Mirror (Read Only)
1

It’s making the general press this hour so of course you already know about it but my question is this: who peers with meta and have you seen BGP sessions drop or the like? Do you operate meta CDN nodes in your network? Are they screaming for help?

This doesn’t sound like it’s a network layer problem but I’m curious.

Cheers,
– jra

2

a message of 124 lines which said:

This doesn't sound like it's a network layer problem but I'm curious.

We can see the start page, authentification fails with an error
message. It does not look like a network issue.

3

They reported a major failure (fat finger source) on one of their DB clusters.

-J

4

My first thought, I'll admit, is that they turned a LLM loose on their
configuration controls, and it "naturally" put itself in route 666.

I sure hope from whatever subbasement clueful admins are still in over
there, they can put it back together. Otherwise, productivity
might improve as users flee to Office 365 to get some work done.

I keep worrying about KEYTRAP....

5

Does meta keep a board somewhere to tell the world it’s down?

6

Does meta keep a board somewhere to tell the world it’s down?

7

This time it looks like AAA issue on Facebook side.

All peerings are up and cdns are reachable

Dominik Dobrowolski

8

Thank you

9

see: Status and outages of Meta business products (metastatus.com)

10

https://metastatus.com/ a message of 210 lines which said:

Does meta keep a board somewhere to tell the world it’s down?

11

Things seem to be returning now, aside from the status page, which
still appears blank. Interesting.

12

Does meta keep a board somewhere to tell the world it’s down?

13

That’s from 2010. You can see if you look at the non-mobile version of the page (requires login though):
https://www.facebook.com/notes/10158791436142200/

image.png

14

What I found intriguing was that I was logged out by Google Docs at the
same moment FB logged me out. Downdetector showed a number of other
supposedly unrelated services with large outage report spikes at roughly
the same time.

mdr

15

Interesting. I have a script that makes a call to TCC every so often to read
temperature data, and it failed from 8:50AM PST to 10:30AM PST.

Ray