Here in Australia these outages happened in in the midde of the night, but generated zapier error alerts as the zaps are contantly pinging Airtable. This caused a lot of concern for our clients, and then time for us (featurepower.com) to investigate.
Hence this post. I couldn’t find any other reference here to this happening. Though I presume many others were affected. So, for info, the errors reported from zapier were:
the app returned “Service Unavailable”
the app returned “503”
Trigger partner failure: the app returned “503”
Trigger partner failure: the app returned “502”
Trigger partner failure: (9001, ‘Max connect timeout reached while reaching hostgroup 20 after 10000ms’)
And also FWIW, zapier turned off some (but not all) impacted zaps because of repeated errors.
I often wonder how we got here, and by “here” I mean about 2 million zaps all running polling API requests against Airtable every few minutes, the vast majority of which result in absolutely nothing to process. And by “here” I also mean a massive Zapier and Integromat framework where no one gave a shred of thought to the idea of integration services that fail, recover, and intelligently catch up gracefully with analytics and advisories that comfort consultants and clients instead of casting fear and uncertainty across their operations.