If they are doing clustering - where not all connections are on one server - they would not have this kind of outage if someone uses an API wrong.
I may be a few years out of the Client/Server space, but load balancing, failovers, and limiting others to impact others was a key to having stability.
And Bob- that sucks. I’m in the middle of a video production- gathering release forms.
Airtable is using AWS for this stuff…
I’m sure they have some concept of auto-scaling servers, failovers, multi-region redundancy, and load-balancing in place, since these are all services offered by, and handled by AWS. But even with the best of setups in a solid service like AWS, scaling software is complicated.
Following our earlier downtime we have implemented a fix and are monitoring the results. Our public API is currently disabled so some services remain unavailable while we continue to monitor this issue.
We expect our public API to be up and running soon, and we will keep you updated on our progress. We deeply apologize for the disruption to your team.
Please let me know if you have any questions or additional issues.
agree @Jeremy_Oglesby - it would not be possible to run a service like Airtable without auto-scaling, clusters, redundancy, etc. My very small agency uses this kind of set up on our client apps and to say that we are a fraction of the size of Airtable is to overstate the case by some margin! So, for sure, the Airtable guys have all of this and more in their toolkit. Doesn’t mean it isn’t frustrating when it goes down, but I know they do this.