Cloudflare went down

The_Right_Honourable_Brit said:
We need to have a serious discussion about the small number of cloud service providers who are seemingly responsible for 99.9% of websites. Eish. So much for resilience/back-up.
Click to expand...
lol had this discussion yesterday with my brother. Like why is it most all website don't have a backup or why is it everyone has stuff on AWS. AWS has been down how many times this past year?
 
HartsockZA said:
lol had this discussion yesterday with my brother. Like why is it most all website don't have a backup or why is it everyone has stuff on AWS. AWS has been down how many times this past year?
Click to expand...
Yeah it's gone down many times this year.
Why don't they have a backup solution?
This has to be affecting many companies
 
The_Right_Honourable_Brit said:
We need to have a serious discussion about the small number of cloud service providers who are seemingly responsible for 99.9% of websites. Eish. So much for resilience/back-up.
Click to expand...

The reality is that 'cloud' hasn't made anything more resilient, it's just concentrated a lot of points of failure. And I am by no means a 'cloud' or infrastructure expert, far from it in fact.
 
HartsockZA said:
lol had this discussion yesterday with my brother. Like why is it most all website don't have a backup or why is it everyone has stuff on AWS. AWS has been down how many times this past year?
Click to expand...

It’s a bit of a misnomer to say AWS has been down X amount of times this year because they run hundreds of services.

They were only really down once where it affected anyone, but **** does down all the time under different services or have maintenance windows etc.

And if people were running their own services the odds are they’d be down far more often, or it would cost them 10x what they pay AWS to achieve the same resiliency.
 
SauRoNZA said:
It’s a bit of a misnomer to say AWS has been down X amount of times this year because they run hundreds of services.

They were only really down once where it affected anyone, but **** does down all the time under different services or have maintenance windows etc.

And if people were running their own services the odds are they’d be down far more often, or it would cost them 10x what they pay AWS to achieve the same resiliency.
Click to expand...
So my point was the down quite often and global outages being rather high for the year, never said anything about the hosting bit or self hosting. I think my point was on often it happens and why probably no one actual has a plan B. But it's cool all is forgiven
 
HartsockZA said:
So my point was the down quite often and global outages being rather high for the year, never said anything about the hosting bit or self hosting. I think my point was on often it happens and why probably no one actual has a plan B. But it's cool all is forgiven
Click to expand...

Every time we consider a Plan B we realise it’s a nice idea but we’ll likely never use it because the effort to make the switch will almost always be longer than the outage.

And I’m referring specifically to these types of outages where you can’t just pop in and change stuff or it’s at a layer you couldn’t multi-host in the first place.

I’m only aware of one AWS outage this year that was really public facing and caused actual downtime.
 
SauRoNZA said:
Every time we consider a Plan B we realise it’s a nice idea but we’ll likely never use it because the effort to make the switch will almost always be longer than the outage.

And I’m referring specifically to these types of outages where you can’t just pop in and change stuff or it’s at a layer you couldn’t multi-host in the first place.

I’m only aware of one AWS outage this year that was really public facing and caused actual downtime.
Click to expand...
Totally agree AWS is far more resilient than most DIY setups.

The catch isn’t uptime, it’s blast radius. When everything funnels through one hyperscale control plane, failures are rare but when they happen, they’re globally visible. That’s the trade-off of centralisation, not a knock on AWS.

To me and you that's nothing... End Customer, the thing is broken and I pay alot of money for **** that fails
 
HartsockZA said:
Totally agree AWS is far more resilient than most DIY setups.

The catch isn’t uptime, it’s blast radius. When everything funnels through one hyperscale control plane, failures are rare but when they happen, they’re globally visible. That’s the trade-off of centralisation, not a knock on AWS.

To me and you that's nothing... End Customer, the thing is broken and I pay alot of money for **** that fails
Click to expand...

Which is why you try to spread the load across distributed locations and running duplicates and auto scale etc hoping that if something does go wrong either the customer doesn’t even notice or it’s at least restricted to a subset of them only.

Does also help to isolate problems, especially when your customers are internal in nature.
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X