Stop raising errors in calculate reactive cache of Environment model

In the Error budget research issue for ~"group::release" we identified ExternalServiceReactiveCachingWorker is one of the top Sidekiq errors.

On further investigation we noticed that all these are Kubernetes pod/cluster unavailability/timeout errors as shown in the below screenshot.

Looks like we are not rescueing these errors in the calculate_reactive_cache method in Enviornment model. In the calculate_reactive_cache we should capture the errors and should return gracefully with a nil or a value which is cacheable instead of throwing exceptions.

Log: https://log.gprd.gitlab.net/goto/6fdc60d0-79e8-11ec-9dd2-93d354bef8e7