Add the ability to collapse diagnostic panels on the cluster page

[antonkovalenko]

Improve cluster page after redesign according ot figma https://nda.ya.ru/t/AAdW9WLO7DUZZy

#2040

Might require additional breakdown into tickets

[adameat]

CPU load is calculated as the cumulative usage across all actor system pools on all nodes in the cluster.

Storage usage is a cumulative usage of raw disk space of all media types.

Memory usage is the total memory consumed by all nodes in the cluster.

Network usage is the average outgoing bandwidth usage across all nodes in the cluster.

cluster statuses (thx to @AlinaBelousovaUX):
🟢 Good = "Everything is working as expected"
🔵 Normal = "All good, some parts of the system are restoring" (we don't expect this status on the cluster status)
🟡 Warning = "There are minor issues"
🟠 Caution = "Critical state, requires immediate attention"
🔴 Failed = "Some systems are failed and not available"
⚪️ Unknown = "The condition cannot be determined" (we don't expect this status on the cluster status either)

[AlinaBelousovaUX]

design review: https://www.figma.com/design/bNbecZuPFDDrJ7ghaXZ2Fe/Tickets?node-id=6927-186025&t=AONqGPGRSo5cuEne-4