Issue #, if available:

resolves #156
resolves #155
resolves #154
resolved #141

Description of changes:

1. Upgraded all versions and cloudformation entries to EKS 1.21.
2. Updated all the tools used by k8s to and eks to latest 1.21.
3. Updated ec2-instance-selector to latest version.
4. Merged the two spot managed groups into the creation of one file so that the creation takes less time that the two executions that we do at the moment.
5. Fixed labels in Self managed section that don’t match anymore with the labels that are created for Managed nodegroups which will break the rest of the workshop.
6. Deployed the node-termination handler to only self managed node groups, the daemonset in this case should only be added to the new instances. Updated page to explain what are the differences and how customers can use the node-termination-handler in queue mode.
7. When doing both the optional self-managed and managed, updated HPA section with increased number of replicas to scale accordingly.
8. Cluster autoscaler in the latest version had OOMKiller running when adding the extra AutoScaling Groups to create and consider. Fixed it with 1 vCPU and 1 GB RAM.
9. Fixed the conflicting names for managed and self-managed node groups.
10. EKS Workshop : Some Instances are not supported in US-WEST-2 #154 - Asked users to stick to the regions which are given in cloud9 creation page. Added a note that as per regions instances could differ and they should filter them out and update node groups files accordingly.

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.