I'd like to know more on how, in a multi-tenant system, do you handle potential conflicts or overlaps in event matching expressions from different tenants? Is that even a thing?
CTO @ Inngest here. Each event is published with an API key associated with a particular account workspace. We add the workspace id along side the each event in the stream so matches are only evaluated against the matching expressions in a particular workspace. This is the first tier of filtering, but many customers send millions of events and have millions of matchers each day so even at the individual workspace level, the optimizations are necessary to prevent backlogs across user accounts.
I wonder how their planning meetings go