Unique target group IDs and health check config per target group

[nickreynke]

Changes made

• Introduce random IDs for target groups to ensure unique naming with replacement on config changes.
• Added validation for the name prefix to enforce a maximum length of 15 characters to not reach limit of target group names.
• Added health check configuration per target group, overriding default target group health check configuration

---

Copilot summary:

This pull request includes several updates to the README.md and main.tf files to enhance the configuration and behavior of the load balancer and its associated resources. The most important changes include adding new requirements and providers, updating the input variables, and improving the configuration of target groups and listeners.

Updates to README.md:

• Added new requirements for null and random providers and listed their versions.
• Included new resources for null_resource and random_id in the resources section.
• Updated the input variables http_ports and https_ports to include more detailed configuration options.
• Specified a maximum length for the name_prefix input variable.
• Added a new input variable for target_group_health_check_protocol.

Enhancements to main.tf:

• Introduced null_resource and random_id resources to manage HTTP and HTTPS target group configurations and ensure unique naming. [1] [2]
• Updated the aws_lb_target_group resources to use the new random_id for unique naming and improved health check configurations. [1] [2]
• Modified the aws_lb_listener resources to directly use values from the configuration without lookup functions for redirect and fixed-response types. [1] [2] [3] [4]

[nickreynke]

Hey @jnonino! Another PR for review :) The only thing I am not sure about is #65 (comment)

[nickreynke]

The new unique naming concept for target groups fixes a bug where Terraform wants to replace the target group with another one with the same name. To make use of the create_before_destroy lifecycle I came up with the random id generation on configuration change. Thus enabling zero-downtime for target-groups.