Added timeout to Postgres queries

[ccollinsuah]

Queries executed by the lambda now have a timeout less than the lambda's max run time, configured by the connection's statement_timeout. Currently, the lambda has at 1 second grace period to cover other operations such as closing the Postgres connection in the event of a query timeout. It may be worth making this grace period either configurable via Terraform or proportional to the lambda's max run time. This flat grace period might not be adequate for S3 writes in cases where a hypothetical long running query returns just shy of the statement_timeout. Let me know if you think either of these additions are appropriate.

Psycopg connections were also not being closed and that has been fixed. Psycopg2 connections use contexts as separate transactions instead of closing at the end of them, so until now the lambda was not gracefully closing its connections.

[camposeddie]

I think everything looks good as a first pass on getting functionality built into the lambda. For the future, it may be better to separate the query timeout from lambda timeout to allow more granular control over each. For our current use case, this works fine 👍