NOAUTH after full queue and reconnect

[mp911de]

com.lambdaworks.redis.protocol.ConnectionWatchdog[] INFO Reconnecting, last destination was 
com.lambdaworks.redis.protocol.CommandHandler[] WARNING [chid=0x17, channel=0xa1ed60d5, ######## -> ] Unexpected exception during request: com.lambdaworks.redis.RedisException: Request queue size exceeded: 100. Commands are not accepted until the queue size drops.
com.lambdaworks.redis.RedisException: Request queue size exceeded: 100. Commands are not accepted until the queue size drops.

[mp911de]

Happens if using Lettuce 4.3 (Ping before connect/calling auth(…)/passing RedisURI with password). Ping before connect auth is fixed in Lettuce 4.4. Calling auth(…) is still broken.

[mp911de]

Increasing the number of allowed commands by the maintenance command limit seems to be a good fit for an intermediate fix. More of a patch, see also #697.