Live uptime updates + inspect API cleanup (Docker connector)

[LordOverlord]

What

This PR improves container uptime handling and modernizes the Docker inspect call:

• Adds a periodic uptime update loop (1s ticker) so uptime increments smoothly without relying on frequent Docker inspect.
• Stores started_at / finished_at timestamps and uses them for live uptime rendering.
• Adds validation/guardrails for bad timestamps and unreasonable uptime ranges; queues a refresh when data is invalid.
• Replaces deprecated InspectContainer usage with InspectContainerWithOptions.
• Initializes uptime meta for new containers and widens the UPTIME column to avoid truncation.

Includes the changes made by @ccarpinteri in this commit:
ccarpinteri@7249d4e#diff-75cd6cc14a79297338822dfe88e22101cfa75b46ec90d8b04d49237305c67ca4

Why:

• Uptime was not updating smoothly (or could get stuck) unless a refresh happened.
• Frequent inspect calls can be expensive and contribute to racey/laggy refresh behavior.
• InspectContainer is deprecated in the Docker client used by this repo.
• UI tweaks ensure uptime is visible and stable.

Notes

• For stopped containers, uptime is displayed as -.
• Refresh requests are queued non-blocking to avoid goroutine/channel stalls.

Testing

• Ran ctop against a local Docker daemon with multiple running/stopped containers.
• Verified uptime increments in real time and remains stable across refreshes.
• Verified no regression in container listing/metadata display.