fix: update units for effectiveCacheSize and tempBuffers

[hrodmn]

⚠️ Checklist if your PR is changing anything else than documentation

• Posted the link to a successful manually triggered deployment workflow (successful including the resources destruction)

Merge request description

The effectiveCacheSize and tempBuffers parameters need to be described in number of 8kb blocks like sharedBuffers.
https://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/Appendix.PostgreSQL.CommonDBATasks.Parameters.html

Successful deployment workflow run: https://github.com/developmentseed/eoapi-cdk/actions/runs/12241535813

[alukach]

This matches what the documentation describes (thanks for linking).

How did you come across this issue? Did this cause you problems?

Is it safe to say that existing eoapi-cdk deployments' effective_cache_size and temp_buffers are thus over-provisioned 8-fold?

[hrodmn]

Thanks for the review @alukach!

How did you come across this issue? Did this cause you problems?

I don't know if it is the source of my issue, but I have been trying to improve performance on the MAAP eoAPI deployment by adding a pgbouncer instance between the titiler-pgstac Lambda and the RDS instance. While trying to determine some good pgbouncer settings for the t3.micro RDS instance, I was looking at the database configuration parameters and saw that effective_cache_size was 6GB for my 1GB instance!

Is it safe to say that existing eoapi-cdk deployments' effective_cache_size and temp_buffers are thus over-provisioned 8-fold?

Yes I think any deployment that is not overriding the default has effective_cache_size that is 8x higher than we want.