AWS backend terraform + Readme #330
Conversation
| count = 1 | ||
| identifier = "${local.name}-writer-${count.index}" | ||
| cluster_identifier = aws_rds_cluster.log_rds.id | ||
| instance_class = "db.r5.large" |
There was a problem hiding this comment.
Worth making this a variable? (maybe in the future/TODO?)
There was a problem hiding this comment.
Yeah, I'm not 100% sold yet on ~all the settings that we should use here, especially for the conformance test. My gut feeling is that we'll stick to Aurora, but maybe we'll move to serverless for conformance testing, but we could also use RDS without Aurora. One interesting data point for conformance testing, is that spinning up or turning down an Aurora instance takes ~5-10 minutes.
There was a problem hiding this comment.
One interesting data point for conformance testing, is that spinning up or turning down an Aurora instance takes ~5-10 minutes
... wow!
Yeah, makes sense to keep all the configs as it for the moment then, can make them variables later once you've decided.
Towards #312
This PR introduces Terraform for AWS storage backend. More work is underway for the other components, see #312.
Putting aside the fact that this runs on AWS (RDS + S3) rather than GCP (Spanner + GCS) , other differences are: