Hacker News new | ask | show | jobs
by L_226 2539 days ago
You can IIRC ping support and ask for a concurrency limit increase, but probably what I would do first is try to segregate lambda deployments and API endpoints (or whatever trigger) by region so that total load is distributed (you get 1000 concurrents per region). Obviously at this point you would also profile your code to optimise function executions.