Is the RAM for a lambda function for one invocation or all concurrent invocations?

Question

As I understand it, AWS Kinesis data stream may have multiple shards. Each shard will invoke one lambda function, and they can happen in parallel.

And so the question is, the RAM that you configure for the lambda function, is it one function or for all functions running concurrently?

For example, if I configure 1GB for my lambda function and 10 lambda functions are invoked concurrently, will the total RAM for all functions be 1GB or 10GB?

Answer 1

RAM is allocated per running instance. Each instance can handle one invocation at a time.

To your question. In total 10GB are allocated each 1GB per instance. Concurrent invocations do not share memory