helm: add wait initContainer and preStop hooks

[hugoShaka]

Part of RFD-096: managing the major upgrades safely

This PR's main purpose is to block proxies running a new Teleport major version from connecting to auth pods running an old Teleport version.

This PR contains three commits:

• one adding initContainers and preStop hooks to the teleport-cluster Helm chart (initContainers were designed in RFD 096, preStop was a nice additoin coming from the wait PR)
• one fixing a bug in the wait command (the DNS error was not properly unwrapped and not recognized as a DNS error)
• one fixing missing override support on some auth Dpeloyment values. As a rule of thumb for future review, we should not use .Values directly and prefer using $auth and $proxy

[webvictim]

I wonder if we could also use this to have proxies unregister themselves on shutdown in future to mitigate #20057

[hugoShaka]

Unregistering would be a nice improvement, but I'm not sure preStop is the best place because it is implementation-specific (Terraform/CloudFormation have the same issue), and we don't have control over the concurrency with preStop. Maybe the proxy will take 2 minutes to shut down, and I don't know what would happen if it gets unregistered while it has not finished doing its job. IMO the best place to do it would be in the graceful shutdown sequence.

Agendas are a little packed right now because of v12, but I'll raise this issue to folks who know more than me about registration and shutdown sequence once things settle down.