Currently if you try to turn on Kafka connector without a working Kafka instance that it can connect to, it doesn't start. This technically makes sense, but the troubleshooting process is made much harder because in order to test some Kafka connector functionality that doesn't pertain to actual Kafka connector operations, you cannot do it without a working Kafka instance.

Current behaviour if Kafka connector cannot connect to a working Kafka instance:

kafka-connector [423198]: ERROR: failed to initialize kafka producer: Failed to create new kafka producer: async producer init failed: kafka: client has run out of available brokers to talk to: read tcp 127.0.0.1:37596->127.0.0.1:8081: i/o timeout.

Expected behaviour should be that Kafka attempts re-connects, instead of shutting down - similar to how Zabbix Server does when database is not available, for example.

423037:20250618:111228.969 [Z3001] connection to database 'zabbix' failed: [2002] Can't connect to local MySQL server through socket '/var/run/mysqld/mysqld.sock' (2)
423037:20250618:111228.969 database is down: reconnecting in 10 seconds