Add naming conventions for infrastructure as code

[psirenny]

I've come across two schools of thoughts for naming resources:

1. Include resource type followed by semantic name — Omit semantic name if one doesn’t exist
2. Exclude resource type from resource name — Provide a semantic name instead; if one doesn’t exist then use this

e.g.:

const rdsCluster = new aws.rds.DbCluster("rds-cluster", {
  engine: "aurora-postgresql",
  engineVersion: "15.4",
  // …
});

new aws.rds.DbInstance("rds-instance-private", {
  dbClusterIdentifier: rdsCluster.id,
  publiclyAccessible: true,
  // …
});

new aws.rds.DbInstance("rds-instance-public", {
  dbClusterIdentifier: rdsCluster.id,
  publiclyAccessible: false,
  // …
});

vs.

const rdsCluster = new aws.rds.DbCluster("this", {
  engine: "aurora-postgresql",
  engineVersion: "15.4",
  // …
});

new aws.rds.DbInstance("private", {
  dbClusterIdentifier: rdsCluster.id,
  publiclyAccessible: true,
  // …
});

new aws.rds.DbInstance("public", {
  dbClusterIdentifier: rdsCluster.id,
  publiclyAccessible: false,
  // …
});

I'm accustomed to approach 1 but it comes with lots of problems. Namely:

• It's verbose
• It's redundant
• It leads to inconsistency because no one agrees on service names (Including the resource Providers themselves 😆)

On the other hand, it probably is more common.