Spent a good chunk of the day dealing with that outage. We were dead in the water and couldn't really do much. Really makes you consider PaaS redundancy but that's usually a huge undertaking. I did find it interesting that Cloudflare has some services that are dependent on some GCP services which exacerbated the problem.
Yeah, it sounds easy but multi-cloud is not easy at all and few do it well. Thing is, when an outage like this happens is when you find the weak points in your systems.
Another reason to run your own stuff if you can (might be "big if" for some), own cloud (e.g. proxmox, k8s) own bitcoin node :-) if you can that is... Then you like the badger, just don't care ...you do you ...
Spent a good chunk of the day dealing with that outage. We were dead in the water and couldn't really do much. Really makes you consider PaaS redundancy but that's usually a huge undertaking. I did find it interesting that Cloudflare has some services that are dependent on some GCP services which exacerbated the problem.
Yeah, it sounds easy but multi-cloud is not easy at all and few do it well. Thing is, when an outage like this happens is when you find the weak points in your systems.
outages of this scale are rare enough that its usually very hard to justify going multicloud given the overhead and cost
Very true. The problem really is there isn't a ton of incentive for these companies to be interoperable with each other.
Yep, we don't use GCP but we do use Cloudflare and we were also impacted
I heard Cloudflare told them a way early and they just sat on it for like an hour... sad
Another reason to run your own stuff if you can (might be "big if" for some), own cloud (e.g. proxmox, k8s) own bitcoin node :-) if you can that is... Then you like the badger, just don't care ...you do you ...
Yeah, running your own infra is best but hard to compete with Google scale.