Infrastructure Apps
Fix Inventory infrastructure apps are low-code applications can be installed to extend Fix Inventory's functionality.
Infrastructure apps can perform simple tasks like cleaning up untagged resources, finding abandoned load balancers, or sending notifications to Slack, Discord, or PagerDuty.
An infrastructure app can not execute any Fix Inventory commands directly. Instead, it generates Fix Inventory commands which are then executed by Fix Inventory.
An app has read-only access to the Fix Inventory Infrastructure Graph using the search()
function, meaning that it can only retrieve information about resources that are already known to Fix Inventory.
To modify resources, an app needs to generate a Fix Inventory command that does so.
Available Apps
cleanup-aws-alarms
This infrastructure app marks all orphaned AWS CloudWatch instance alarms (e.g., alarms associated with an EC2 instance that no longer exists) for cleanup.
cleanup-aws-loadbalancers
This infrastructure app cleans up AWS ALB/ELB load balancers with no instances attached to them.
cleanup-aws-vpcs
When a VPC is marked for cleanup, this infrastructure app marks its network resource dependencies for cleanup as well.
cleanup-expired
This infrastructure app looks for resources with the tags expiration or fix:expires and flags them for cleanup if they are expired.
cleanup-untagged
This infrastructure app deletes cloud resources that are missing mandatory tags after a defined amount of time has passed since their creation.
cleanup-volumes
This infrastructure app cleans up unused storage volumes.
protector
This infrastructure app protects important resources from deletion by Fix Inventory.
tagvalidator
This infrastructure app validates the contents of expiration tags. With it you can enforce a max. expiration length for certain resources in an account. For instance you could have an org policy that says in our "dev" account compute instances are only allowed to exist for 2 days max. Then this infrastructure app can ensure that the expiration tag on those instances is set to no more than 2 days. If it is set to e.g. 50h it would be corrected down to 48h.