HA does not stop retrying after 5 minutes

[lippserd]

There are two problems that make HA retry infinitely often:

• If the instance is not responsible, there is no retry timeout. Therefore, only errors that occur during the connection setup are retried for a maximum of 5 minutes.
• If the instance is responsible, there is also no timeout for retrying, but a deadline context of 1 minute is used that ends earlier than our default timeout for retries of 5 minutes, and thus renders it ineffective.