updated cloudtrail module

modules/cloudtrail/README.md

@@ -4,4 +4,18 @@
 - All management and global events are captured within CloudTrail
 - Store CloudTrail logs are stored in a private s3 bucket
 - Access to the above bucket should also be logged by CloudTrail
-- Integrate 
+- Integrate CloudTrail with CloudWatch Logs
+- Ensure CloudTrail logs are encrypted at rest using KMS customer managed keys (CMKs)
+- Enable CloudTrail log file integrity validation
+
+# Notes:
+- Create s3 private bucket 
+- logging to own bucket doesn't work, can create another bucket to save logs to and add those logs to the cloudtrail bucket? 
+- created kms cmk in the console, which can be imported into our config and used to encrypt/decrypt logs. 
+- bucket permissions may cause issues later, `private` vs `log-delivery-write`
+    - private = owner has FULL_CONTROL permissions
+    - log-delivery-write = LogDelivery group gets WRITE and READ_ACP permissions
+        - https://blog.runpanther.io/s3-bucket-security/
+        - based on this^ it seems like log-delivery-write is secure enough
+- as per gov.uk website, companies must keep records for 6 years from the end of the financial year they relate to, following the same rules to keep logs for 6 years after creation.  `days = 2190`
+    - need to confirm that this is 6 years after input not 6 years after bucket creation

modules/cloudtrail/main.tf

@@ -0,0 +1,82 @@
+# for now we are importing an existing key 
+# this key enables encrption/decryption of logs
+resource "aws_kms_key" "bucket_key" {
+    description = "key for cloudtrail logs bucket"
+
+    # policy here, by default key has all permissions if a policy is not defined 
+    # a policy was already defined when I created the key in console.
+}
+
+resource "aws_s3_bucket" "cloudtrail_logs" {
+    bucket = "cloudtrail_logs"
+    acl = "log-delivery-write" 
+
+    logging {
+        # logging to own bucket doesn't work
+        target_bucket = aws_s3_bucket.cloudtrail_bucket_logs.id
+    }
+
+    server_side_encryption_configuration {
+        rule {
+            apply_server_side_encryption_by_default {
+                kms_master_key_id = aws_kms_key.bucket_key.arn
+                sse_algorithm = "aws:kms"
+            }
+        }
+    }
+
+    # versioning ensures that when the bucket recieves multiple requests at the same time
+    # it stores all of them, this makes sense to enable for logging
+    versioning {
+        enabled = true
+    }
+
+    expiration {
+        days = 2190
+    }
+}
+
+# there is probably a better way to do this 
+# this bucket exists so that cloudtrail bucket logs can be sent here
+# the plan is to route these logs back into the cloudtrail bucket somehow
+resource "aws_s3_bucket" "cloudtrail_bucket_logs" {
+    bucket = "cloudtrail_bucket_logs"
+    acl = "log-delivery-write"
+    logging {
+        target_bucket = aws_s3_bucket.cloudtrail_logs.id
+    }
+
+    server_side_encryption_configuration {
+        rule {
+            apply_server_side_encryption_by_default {
+                kms_master_key_id = aws_kms_key.bucket_key.arn
+                sse_algorithm = "aws:kms"
+            }
+        }
+    }
+
+    versioning {
+        enabled = true 
+    }
+
+    expiration {
+        days = 2190
+    }
+}
+
+resource "aws_cloudtrail" "cloudtrail_logging" {
+    name = "cloudtrail_logging" 
+    s3_bucket_name = aws_s3_bucket.cloudtrail_logs.id
+    include_global_service_events = true
+    enable_log_file_validation = true 
+    is_multi_region_trail = true 
+    is_organisation_trail = true
+    kms_key_id = aws_kms_key.bucket_key.arn
+
+
+    # events to track 
+    event_selector {
+        include_management_events = true 
+    }
+
+}

modules/default_vpc/README.md

@@ -13,7 +13,7 @@ After creating a new IAM user, named 'test_2', in the AWS Consule. I logged on a
 
 **The aim of the policies in this module is that, when attached to an IAM user, they would remove the default VPN and all its associated infrastructure.**
 
-### Thought Process:
+### Notes:
 - It is hard for Terraform to remove a VPC that it did not create (you have to trick it into thinking it did)
 - tried importing the vpc into terraform then applying with `count = 0` to destroy, which returned success message but didn't actually delete the vpc (checked using AWS console)
 - tried importing the vpc then `terraform destroy` which also returned a success message but didn't delete the vpc 
@@ -25,4 +25,7 @@ After creating a new IAM user, named 'test_2', in the AWS Consule. I logged on a
 - succesfully deleted vpc and all associated infrastructure in one region (`us-east-2`)
 - tested in second region to confirm that only the subnets, vpc, and igw are required for import. 
 - Moved terraform script into the default_vpc directory, create python script to loop through regions, can use `boto3` to get vpc_ids, subnet_ids, subnet_cidr and igw_id
-- tried to condense subnet resources into one with `count` but didn't work because then it requires `terraform destroy` to delete the subnet. 
+- tried to condense subnet resources into one with `count` but didn't work because then it requires `terraform destroy` to delete the subnet. 
+- look up terraform `where` clause to explicitly delete subnets / vpc where it's the default, as opposed to filtering in python. 
+- test whether subnet cidr is necessary or if can be left blank (`"0.0.0.0/0"`)
+    - if it doesn't need to be correct, it makes the python script much easier