I recently ran into a scenario where using serverless caused a cascading failure. We have database searches that are executed directly via AWS lambdas. When we

I recently ran into a scenario where using serverless caused a cascading failure. We have database searches that are executed directly via AWS lambdas. When we sent too many queries in too short of a time that took down the database and which then took down other things. (The fact multiple things are connected to this database is another thing to solve.) While the database was down the lambdas were still sending more and more requests. If this was a standard service the standard service would have became overloaded very quickly, allowing the database to remain up but at a degraded performance and the only service that would have been down would have been the one that was getting too many queries. Instead, everything was effected. I've came to the realisation that if your lambda requires something else to operate, lambda is probably not a good solution.