[ENHANCEMENT] ring/backoff: reuse timers in lifecycler and backoff loops

[sandy2008]

What this PR does:

• Replaces repeated time.After(...) calls with reusable time.Timer instances in:
  • pkg/ring/lifecycler.go
  • pkg/ring/basic_lifecycler.go
  • pkg/util/backoff/backoff.go
• Adds shared safe timer utilities (stopAndDrainTimer, resetTimer) in pkg/ring/ticker.go to avoid timer channel races/leaks.
• Optimizes DynamoDB CAS by preallocating putRequests map capacity in pkg/ring/kv/dynamodb/client.go.
• Keeps behavior unchanged while reducing allocations in hot loops.

Performance validation (local -benchmem):

• Watch loop (time.After vs reusable timer): 248 B/op, 3 allocs/op -> 0 B/op, 0 allocs/op.
• Backoff wait (time.After style vs reusable timer): 248 B/op, 3 allocs/op -> 0 B/op, 0 allocs/op.
• Observed small ns/op improvement in both benchmarks.

Which issue(s) this PR fixes:
Similar to #7266

Checklist

• Tests updated
• Documentation added
• CHANGELOG.md updated - the order of entries should be [CHANGE], [FEATURE], [ENHANCEMENT], [BUGFIX]

[friedrichg]

I am concerned also that unit tests on backoff.go don't cover these changes.
could you add them?

[sandy2008]

@friedrichg Oops.. forgot to modernize, let me fix.

[friedrichg]

LGTM. Thank you!