[P2][Logs] Reclassify expected events to lower log levels

Parent issue: #4691

Intent

Expected events such as normal retries, duplicate ready messages, stale epoch drops, and routine shutdowns are currently too visible at Info or Warn. This dilutes the signal of logs that actually require action. This task reclassifies expected events so default levels emphasize actionable conditions.

Modification points

• Review the targeted paths and identify events that are expected in steady-state or during routine recovery.
• Move those events to Debug or sampled Info when they are useful for diagnosis but not actionable by default.
• Keep Warn and Error for events that indicate a real anomaly, user-visible impact, or operator action.

Entry points

• logservice/logpuller/region_request_worker.go:214
• downstreamadapter/eventcollector/dispatcher_stat.go:615
• downstreamadapter/eventcollector/dispatcher_stat.go:681
• server/module_election.go:182
• pkg/etcd/client.go:286

Done criteria

• Default Warn logs are closer to events that need action.
• Default Info logs focus on important phase changes and diagnostic conclusions.
• Expected routine events are still observable when debug or sampling is enabled.

Start by reviewing the listed entry points: logservice/logpuller/region_request_worker.go:214, downstreamadapter/eventcollector/dispatcher_stat.go:615 and :681, server/module_election.go:182, and pkg/etcd/client.go:286. Identify routine events and adjust their levels or sampling so Warn and Error indicate actionable conditions while expected events remain visible through debug or sampling.