Paws auth failures when running in a contanerized Lambda

[glennlawyer]

As of Dec 2020, AWS Lambda now supports deploying AWS Lambda functions as container images. This allows us to run R code in Lambda.

I've build a docker image involving some R code. The R code needs to access a database hosted on AWS, so I've installed the PAWS package. I've given the Lambda the proper IAM permissions to access the DB.

However,

rds <- paws::rds()
token <- rds$build_auth_token(endpoint=paste0(settings$server, ":", settings$port), 
						  user=settings$user)

fails with the message

"Error in system(sprintf(\"echo $%s\", var), intern = T): cannot popen 'echo $AWS_CREDENTIAL_EXPIRATION', probable reason 'Cannot allocate memory'

Is there a workaround?

[glennlawyer]

Issue resolved. I allocated more memory to the Lambda.

In the SAM template:

MemorySize: 1024