Auto-generated
This page is generated by terraform-docs at build time from the infra/ repo. Do not edit manually.
AWS Backup plan¶
One backup vault, one plan and one explicit selection per environment, with the IAM role AWS Backup assumes to take and restore the backups. Part of the account security baseline (IS-120, follow-up to IS-67).
Two rules, both at the same hour:
daily, keptdaily_retention_daysmonthly, on the 1st, keptmonthly_retention_days
On the 1st both rules fall in the same window; AWS Backup then takes a single backup and keeps it for the longer of the two retentions.
Resources are selected by ARN, never by tag, so nothing joins the plan by accident. No cold storage: the EBS archive tier has a 90-day minimum and 24-72 h restores, which does not fit short retentions.
The vault uses the AWS-managed aws/backup key. EBS and RDS recovery points
keep the source resource's own encryption key regardless, so the backup role
needs those keys to allow IAM-delegated access (each key policy's root
statement) or to name the role explicitly.
When alarm_actions is set, a CloudWatch alarm fires on any failed, aborted
or expired job in the vault: the plan this replaces failed every day for 19
months without anyone noticing.
Requirements¶
| Name | Version |
|---|---|
| terraform | >= 1.6.0 |
| aws | ~> 6.0 |
Providers¶
| Name | Version |
|---|---|
| aws | ~> 6.0 |
Modules¶
No modules.
Resources¶
| Name | Type |
|---|---|
| aws_backup_plan.this | resource |
| aws_backup_selection.this | resource |
| aws_backup_vault.this | resource |
| aws_cloudwatch_metric_alarm.jobs_failed | resource |
| aws_iam_role.this | resource |
| aws_iam_role_policy_attachment.backup | resource |
| aws_iam_role_policy_attachment.restore | resource |
| aws_iam_policy_document.assume_role | data source |
Inputs¶
| Name | Description | Type | Default | Required |
|---|---|---|---|---|
| alarm_actions | SNS topic ARNs notified when a job fails, aborts or expires. Empty creates no alarm. | list(string) |
[] |
no |
| base_tags | Base tags for all resources. | map(string) |
{} |
no |
| completion_window_minutes | Minutes a job may run before AWS Backup marks it expired. Must cover the largest first (full) snapshot. | number |
720 |
no |
| daily_retention_days | Days each daily recovery point is kept. | number |
14 |
no |
| monthly_retention_days | Days each monthly recovery point (taken on the 1st) is kept. 60 keeps the two most recent monthlies. | number |
60 |
no |
| name | Name of the plan and selection, and prefix for the vault (<name>-vault), role (<name>-role) and alarm. |
string |
n/a | yes |
| resource_arns | ARNs of the resources to back up (EBS volumes, RDS instances ...). Explicit by design: nothing is selected by tag. | list(string) |
n/a | yes |
| start_hour_utc | Hour (UTC, 0-23) the daily and monthly jobs start. Keep it clear of the RDS backup and maintenance windows of the selected databases. | number |
1 |
no |
Outputs¶
| Name | Description |
|---|---|
| plan_id | ID of the backup plan. |
| role_arn | ARN of the IAM role AWS Backup assumes. Pass it to modules whose KMS key policy grants the backup role explicitly (e.g. the vault ClickHouse module's backup_role_arn). |
| vault_name | Name of the backup vault. |