Allow configuration of the leaked ENI reconciliation

[alanty]

What would you like to be added:
As of v1.12.6 the CNI checks for "leaked" ENIs every hour:
https://github.com/aws/amazon-vpc-cni-k8s/blob/v1.12.6/pkg/awsutils/awsutils.go#L440

	// Clean up leaked ENIs in the background
	if !disableENIProvisioning {
		go wait.Forever(cache.cleanUpLeakedENIs, time.Hour)
	}

rather than having the time.hour hardcoded could we parameterize that interval, or allow this process to only run once on node init?

Why is this needed:

This check eventually issues a DescribeNetworkInterfaces command to get all available ENIs that were allocated by the AWS CNI plugin, but were not deleted.
https://github.com/aws/amazon-vpc-cni-k8s/blob/v1.12.6/pkg/awsutils/awsutils.go#L1748

In accounts with a large number of nodes or network interfaces this can cause a large number of Describe calls. By controlling the reconciliation we can continue to allow the leaked ENIs to be cleaned up, but reduce the unnecessary Describe calls for large accounts.

[jayanthvn]

@alanty - We are thinking about moving this to the control plane (VPC RC) hence planning to add this knob - #1641 but that would need some additional testing and release so in the mean time we can think of making it configurable or make it staggered. It will be still suboptimal even if it is called once since all nodes will make a call once and if there is a node churn then we won't get much of the benefit.

[jdn5126]

Closing as fixed by #2370

[github-actions]

⚠️COMMENT VISIBILITY WARNING⚠️

Comments on closed issues are hard for our team to see.
If you need more assistance, please open a new issue that references this one.
If you wish to keep having a conversation with other community members under this issue feel free to do so.