Interestingly, Uptime robot indicates one of my websites is down, but I can reach it myself from my residential network as well as from a bunch of other places using a VPN.
@stevewatson301 said:
Interestingly, Uptime robot indicates one of my websites is down, but I can reach it myself from my residential network as well as from a bunch of other places using a VPN.
Woke up in the middle of the night due to monitoring alerts
All servers are still online though, wtf
realize it's Cloudflare, attempting to implement a fix
Trying to test a fix but needed to edit CDN settings. Long loading times and search doesn't load results on Cloudflare's dashboard. Everyone's probably slamming CF's website
Keep trying but service came back before I could test the fix
Even though I have redundancy built in, longest service disruption for my API since inception
(Global) Outages like that always makes me wonder how monitoring tools like HetrixTools (tagging @HBAndrei ) handles that.
How does the load goes up? How many notifications? How much traffic spike? Many people logging for first time in months to check wtf?
--
Like one of the "service down" e-mail showed up in my inbox (Created on: 21 June 2022 at 08:32 (Delivered after 5604 seconds)) 90 minutes after it was sent and with GMail sorting it's like 10 e-mails ABOVE "service UP" e-mail.,
EDIT: Oh, yeah. That account e-mails goes via Cloudflare catch-all, I guess that kinda explains why it was 90 minutes later. Outageception
@JabJab said:
(Global) Outages like that always makes me wonder how monitoring tools like HetrixTools (tagging @HBAndrei ) handles that.
How does the load goes up? How many notifications? How much traffic spike? Many people logging for first time in months to check wtf?
--
Like one of the "service down" e-mail showed up in my inbox (Created on: 21 June 2022 at 08:32 (Delivered after 5604 seconds)) 90 minutes after it was sent and with GMail sorting it's like 10 e-mails ABOVE "service UP" e-mail.,
EDIT: Oh, yeah. That account e-mails goes via Cloudflare catch-all, I guess that kinda explains why it was 90 minutes later. Outageception
It was quite the event indeed. "Luckily" for us, it wasn't the first time, and we've learned a lot from previous similar experiences.
Emails sent out spiked quite a lot, to over 10 times the usual numbers:
The detected errors spiked to about 7 times the usual numbers:
There was indeed a traffic spike on the platform, as is to be expected when so many websites go down all at once:
Comments
Its okay now, things are going back up.
That why I cant access hostcram lately
xVideos still up, what a stud
Up for me now.
that is why I am unable to access LET and other forums too.
It is @tinyweasel fault!!! He is DDoSing Internet!
Will be interesting to see the RFO. Will probably be available in a few hours I'd think.
Getting too frequent which doesn't look on company this large.
It seems to have returned to normal.
apolgy for bad english
where were you when cloudflare die
i was at house eating dorito when phone ring
“cloudflare is kil”
“no”
May be related to recent publication of ddos origins? 😀🤣
involucrated?
omgwtfcf
Yes, Cloudflare users are definitely involucrated in this unfortunate incident.
Interestingly, Uptime robot indicates one of my websites is down, but I can reach it myself from my residential network as well as from a bunch of other places using a VPN.
Looks like a partial outage of sorts.
Partial, and intermittent apparently.
Just waiting for the RFO/RCA
Woke up in the middle of the night due to monitoring alerts
All servers are still online though, wtf
realize it's Cloudflare, attempting to implement a fix
Trying to test a fix but needed to edit CDN settings. Long loading times and search doesn't load results on Cloudflare's dashboard. Everyone's probably slamming CF's website
Keep trying but service came back before I could test the fix
Even though I have redundancy built in, longest service disruption for my API since inception
FeelsBadMan.png
(Global) Outages like that always makes me wonder how monitoring tools like HetrixTools (tagging @HBAndrei ) handles that.
How does the load goes up? How many notifications? How much traffic spike? Many people logging for first time in months to check wtf?
--
Like one of the "service down" e-mail showed up in my inbox (Created on: 21 June 2022 at 08:32 (Delivered after 5604 seconds)) 90 minutes after it was sent and with GMail sorting it's like 10 e-mails ABOVE "service UP" e-mail.,
EDIT: Oh, yeah. That account e-mails goes via Cloudflare catch-all, I guess that kinda explains why it was 90 minutes later. Outageception
Unbelievable!
waiting for their postmortem/RCA...
https://blog.cloudflare.com/cloudflare-outage-on-june-21-2022/
Defund BGP!
Route Packages manually!
Added this to the OP.
It was quite the event indeed. "Luckily" for us, it wasn't the first time, and we've learned a lot from previous similar experiences.
Emails sent out spiked quite a lot, to over 10 times the usual numbers:

The detected errors spiked to about 7 times the usual numbers:

There was indeed a traffic spike on the platform, as is to be expected when so many websites go down all at once:

You can read our full description of the outage, here:
https://docs.hetrixtools.com/june-21-2022-cloudflare-outage/
I think it was handled OK, but there's always room to improve.
Cheers.
I was asleep.
What did I miss?
hmmmm thats why i not use cloudflare , if cloudflare goes down so as your website too.
Well that's not entirely accurate.
If Cloudflare goes down, your website doesn't "go down".
It's hosted elsewhere which is online and where you have other ways of accessing - just won't be available/accessible by the public
This issue took LowEndTalk down !!!