Updating the template

The current template version is 1. We support the current version and the one before it at the same time (now: 1), so an update is never urgent. An update never replaces a table or a bucket: a release that would do that fails our checks before it’s published.

Know when there’s an update

When a new version is out, the console’s Setup page asks you to update, and the template_version check in /health shows template_outdated.

curl "$BASE/health" -H "Authorization: Bearer $KEY"

Update the stack

1. Sign in to the console and open Setup.
2. Follow the update link. If you subscribed through AWS Marketplace, this opens Marketplace’s guided install. Otherwise it opens a CloudFormation update with the new template and your current parameters.
3. Review the change set. It should show changes to policies, rules or alarms, never a replacement of a table or bucket.
4. Tick the IAM acknowledgement and click Update stack.

Check it worked: the stack reaches UPDATE_COMPLETE, and the template_version check shows "ok": true.

Change a parameter

To turn on the CloudTrail trail, data events, or your own KMS key, update the stack with the same template and new parameter values. For example, to create the trail:

aws cloudformation update-stack --stack-name rhino-relay --use-previous-template \
  --capabilities CAPABILITY_NAMED_IAM \
  --parameters ParameterKey=TenantId,UsePreviousValue=true ParameterKey=ExternalId,UsePreviousValue=true \
    ParameterKey=SendingDomain,UsePreviousValue=true ParameterKey=ArchiveRetentionDays,UsePreviousValue=true \
    ParameterKey=KmsKeyArn,UsePreviousValue=true ParameterKey=CreateTrail,ParameterValue=true \
    ParameterKey=DataEvents,UsePreviousValue=true ParameterKey=Disconnect,UsePreviousValue=true

Use your stack’s name if it isn’t rhino-relay. The parameters are listed in What the template creates.

Check it worked: the stack reaches UPDATE_COMPLETE and its resources include Trail and TrailBucket.

If an update fails

CloudFormation rolls a failed update back by itself. If the rollback fails too, the stack shows UPDATE_ROLLBACK_FAILED. Continue the rollback, skipping the resource that failed:

aws cloudformation continue-update-rollback --stack-name rhino-relay --resources-to-skip <LogicalId>

Tables and buckets are kept on every path, so your data is safe either way. If you’re stuck, contact support with the stack’s events.