You might be running into something related to cold starts: https://blog.octo.com/en/cold-start-warm-start-with-aws-lambda/, https://epsagon.com/blog/how-to-minimize-aws-lambda-cold-starts/

Yeah, I was thinking it was something along those lines. Is that a common thing for ions? I haven't seen anyone specifically discussing cold starts with ions, so I wasn't sure if I was doing something weird or if it was expected behavior. Thanks for the links!

Sorry, I don't use Ions in particular. Many AWS Lambda users use hacks to keep their Lambdas warm -- often that is managed by the serverless framework you're using (https://serverless.com/blog/keep-your-lambdas-warm/). I wonder if Ions does or ought to do something similar

Yeah, I had this impression that ions had a way around the cold start. Thanks for the help!

@U054BUGT4 What we have done at work is create a cloudwatch event that passes some json to the lambda function to keep it warm once per minute. It’s pretty trivial to implement.

Perfect, that sounds like exactly what I need. I thought I might be missing some obvious datomic specific thing (or that my ion had something wrong with how I set it up. Thanks!

I'm running across a similar issue this afternoon when using Omniconf to populate our config. I'm not seeing a "too many results" error, though one could be masked by the Omniconf SSM code: we're just not getting values populated from SSM. In our case, too, nothing has changed in the config either in the SSM parameter store or in our config-calling code. In our case we're in us-west-2.

Interesting. Do you have a reference for this issue? Where this limit is documented?

Fixed the Omniconf issue. The AWS API returns 10 parameters and a NextToken value when there are more. https://github.com/Dept24c/omniconf/blob/ssm-recursive-next-token/src/omniconf/ssm.clj