Just deployed our first lambdas into production, and we are finding a few lambda errors when there are traffic spikes, but we cant find them in cloudwatch logs, which makes us wonder if they happen when the lambdas are starting.
Does lambda log all errors into logs in cloudwatch? Is there a way to find these errors caused by spikes, we think they happen when the lambdas are scalling up, since we from 40-(+4000) invocations in a given time? i have tried all sorts of metrics and i cant find the buggers, they dont seem to caused any logic in our code either (as we have all sorts of metrics for these).
Is there a way to find them?