terminationGracePeriodSeconds not configurable for the sidekiq chart

Summary

Currently the terminationGracePeriodSeconds is not a configurable option for the sidekiq chart.

Reference: https://gitlab.com/gitlab-org/charts/gitlab/-/blob/e47e8f8d5178f9ebaafd9ba8f6a42644e2999e47/charts/gitlab/charts/sidekiq/templates/deployment.yaml

This limits a deployment to forcibly removing a Pod after 30 seconds. For the most part, this should be fine for smaller installations. For larger ones, we may need to be more forgiving if some sidekiq jobs take a long time, or if there's a problem with sidekiq's ability to requeue a job during it's termination period. GitLab.com would like to configure the SIDEKIQ_TIMEOUT to a larger value to reduce errors on jobs that typically run longer than the default 5 seconds. Increasing this timeout too greatly will interfere with Sidekiq's ability to perform the necessary options in time if set to a value greater than terminationGracePeriodSeconds.

Due to the ability to specify differing types of jobs on differing sidekiq deployments, it would be beneficial to ensure that we can configure this grace period per Pod, or sidekiq deployment.

Considerations

• Allow terminationGracePeriodSeconds to be configurable per Pod
• Protect installations where the sidekiq.timeout is set to a higher value than the default or configured value of terminationGracePeriodSeconds