WI personaly do, and the downtime is the exact reason why, I have an automatic banner when a third party is down that tells my customers X feature is down because it relies on Y which is currently unavailable, sorry blablabla". I find this way better than saying "oh yeah X feature is down we're working on it" when I can't do anything about it
Some platform will claim to release "native feature" that in fact relies on someone else's work without disclosing it.
But I prefer being transparent with my users. Also if you sign some big companies there is a whole process you have to go through where you disclose everything, if I don't know what my third party relies on its hard to be completely honest in those kind of process.
in EU, some customers will chimp out if they hear the word "google" or "amazon".
I don't blame CF, I just dont like how they word absolutely everything like they made the whole infra from scratch when its not the case, look at their status page, they say they're actively working on it, but if its a GCP issue, what are they really doing ? making some phone calls maybe
Some platform will claim to release "native feature" that in fact relies on someone else's work without disclosing it.
But I prefer being transparent with my users. Also if you sign some big companies there is a whole process you have to go through where you disclose everything, if I don't know what my third party relies on its hard to be completely honest in those kind of process.
in EU, some customers will chimp out if they hear the word "google" or "amazon".
I don't blame CF, I just dont like how they word absolutely everything like they made the whole infra from scratch when its not the case, look at their status page, they say they're actively working on it, but if its a GCP issue, what are they really doing ? making some phone calls maybe
WNo but maybe they can say in their documentation when a feature relies on a third party and which one, just a little bit of transparency you know.
WAlso I'm pretty sure they would be able to do everything self hosted, im not sure they could without updating plans pricing though
They already displayed they were able to do some crazy stuff
They already displayed they were able to do some crazy stuff
JI'm one of the first to critique Cloudflare's marketing - it's bad for many other reasons. But basically everything is built on top of other parts, especially online. It's unreasonable to expect third party usage disclaimers on everything - should they also include the OSS software they use, and versions of it, in case they upgrade to a new version and that breaks? Or their DB solutions and how/where they host them?
It's just not a fair expectation. Outages suck, and I can promise you Cloudflare feel that a lot - these kinds of things impact their own tooling, internal usage, and bottom line, but we'll get a post-mortem in the coming days I imagine, and they'll be talking about all the things they're doing to mitigate this moving forward.
It's just not a fair expectation. Outages suck, and I can promise you Cloudflare feel that a lot - these kinds of things impact their own tooling, internal usage, and bottom line, but we'll get a post-mortem in the coming days I imagine, and they'll be talking about all the things they're doing to mitigate this moving forward.
JAs Dane said on Twitter recently, at the end of the day, the outage is Cloudflare's responsibility; https://x.com/dok2001/status/1933272215396688124. The vast majority of people don't need (or want) to know the implementation details like vendors used, software, hosting locations, etc. and that's fine. And if you're someone that does, there's tonnes of people around here and elsewhere on the web who'll happily chat about them 
JI mean.. if you thought you had your redundancy sorted by deploying to GCP as well as Cloudflare, and then this happened, you might be a little miffed
Jfor example I remember it always being common knowledge that Heroku ran on AWS
Jwhich allowed you to make sensible decisions regarding deploying to multiple clouds for redundancy
WI totally get your point, and i'm crossing fingers that the postmortem will include the keyword "google" or "gcp". I didn't see that tweet from Dane and its a nice reaction from him, but the user asked if it has something to do with google and he didn't answer, I just find this a bit frustrating, and maybe the postmortem will answer those questions. My whole point is that I don't care if CF relies on third parties, I would just want to know when they do.
I left Azure a few years ago, even though I had thousands in credits, because I had uptime issue all the time for frontends. When I make my decision on which cloud i'd like to go with, knowing they could eventualy depend on another provider I was very disappointed with could be great. (e.g: if CF relied on Azure for pages I would have not go with them)
As for the OSS software, in my own postmortem I do mention when I upgraded something and it broke to my customers because some are developpers/devops (using my APIs). When they read all that, they're always much more receptive, and they can also inform their colleagues or customers about what exactly happened instead of just saying "yeah it stopped working for some reason"
I left Azure a few years ago, even though I had thousands in credits, because I had uptime issue all the time for frontends. When I make my decision on which cloud i'd like to go with, knowing they could eventualy depend on another provider I was very disappointed with could be great. (e.g: if CF relied on Azure for pages I would have not go with them)
As for the OSS software, in my own postmortem I do mention when I upgraded something and it broke to my customers because some are developpers/devops (using my APIs). When they read all that, they're always much more receptive, and they can also inform their colleagues or customers about what exactly happened instead of just saying "yeah it stopped working for some reason"
DAfaik they already have their own servers...
