optimize: merge channel availability log output and adjust log level

[shenshichao163-oss]

What this PR does

Optimizes the channel availability thread log output in NettyClientChannelManager.

Changes

1. info level: no stack traces — Changed LOGGER.error() to LOGGER.info() for the summary log, which outputs all failed addresses in a single entry without stack traces.
2. debug level: detailed stack traces — Changed LOGGER.error() to LOGGER.debug() for per-address detailed logs with full stack traces.
3. Merged output — All failed addresses are output in one log entry at info level (using failedMap.keySet()), instead of one log per failed address.
4. Independent log levels — Changed else if to two independent if blocks, so info and debug levels work independently rather than being mutually exclusive.

Issue Link

Fixes #5474

Before

• LOGGER.isInfoEnabled() → LOGGER.error() (wrong level, outputs error-level logs when info is enabled)
• else if LOGGER.isDebugEnabled() → LOGGER.error() (only reached when info is disabled)
• Log level mismatch: info branch uses error, debug branch uses error

After

• LOGGER.isInfoEnabled() → LOGGER.info() (summary, no stack traces)
• LOGGER.isDebugEnabled() → LOGGER.debug() (per-address detailed stack traces)
• Both levels work independently

Checklist

• I have searched the issues of this repository and believe that this is not a duplicate.
• I have checked that this modification does not cause CI to fail.
• I will sign the Apache CLA if required.

[WangzJi]

Thanks for the fix. The log level changes make sense, but #5474 is not fully addressed yet. doConnect() still logs the exception at ERROR before doReconnect() aggregates it, so an INFO-level client still prints one stack trace per unavailable address. Could you handle this path as well and keep stack traces only at DEBUG?