Provide possibility for Nginx cache invalidation after staging

[philippsauer]

Is your feature request related to a problem? If yes, please describe it.

In order to reach certain performance goals, we are forced to increase the nginx cache time to 24h. That works well but has downsides: Despite a nightly staging, changes in the CMS are not reflected immediately in the cache and can take up to 24h to become visible, depending on when the related cache entry was created. Unfortunately, this is not acceptable for the customer.

Describe the desired solution.

We want to be able to (automatically) trigger nginx cache invalidation without involving Dev or a DevOps Engineer.

1. Ideally, there would be two modes: full or channel-specific. Currently, the cache keys are hashed which would make this difficult, maybe they could be prefixed with an identifier that points to a certain channel?
2. Also, it should be possible to somehow automate this invalidation. (E.g. via a scheduled job etc.)
3. Bonus: The invalidation can be triggered via ICM, like the channel page cache

Provide additional context.

FYI: There is a related feature request in the Helm chart repository: During Deployments, the cache needs to be invalidated manually by restarting the Memcached container to prevent broken pages, see Helm Chart Issue: https://github.com/intershop/helm-charts/issues/410

AB#89722