fix(ci): wait for the kafka broker before creating topics

[AlinsRan]

Description

t/plugin/kafka-logger.t TEST 18 fails intermittently in CI, e.g. https://github.com/apache/apisix/actions/runs/29567552409/job/88251295446:

#   Failed test 't/plugin/kafka-logger.t TEST 18: report log to kafka by different partitions - pattern "(?^:partition_id: 1)" should match a line in error.log (req 0)'
#   Failed test 't/plugin/kafka-logger.t TEST 18: report log to kafka by different partitions - pattern "(?^:partition_id: 2)" should match a line in error.log (req 0)'

The cause is in the CI setup, not in the test. Earlier in the same job log:

Error while executing topic command : Replication factor: 1 larger than available brokers: 0.
Error while executing topic command : Replication factor: 1 larger than available brokers: 0.
Created topic test4.

ci/init-plugin-test-service.sh creates the topics right after make ci-env-up returns. When the broker has not registered itself in ZooKeeper yet, creating test2 and test3 fails. The failure is ignored, and since KAFKA_CFG_AUTO_CREATE_TOPICS_ENABLE=true, test3 is then auto-created with the broker default of one partition on first produce.

TEST 18 expects test3 to have 3 partitions and asserts that partition ids 0, 1 and 2 all show up in the error log. With a single-partition topic every message goes to partition 0, so partition_id: 1 and partition_id: 2 never appear — exactly the observed failure. Whether the run is affected depends purely on broker startup timing, which is why it looks flaky.

Fix: retry topic creation until the broker accepts it, and fail the setup loudly instead of starting a test run against a wrong partition layout. ci/init-last-test-service.sh had the same pattern and gets the same treatment.

Verified locally by reproducing the race (broker deliberately started 20s late):

Error while executing topic command : Replication factor: 1 larger than available brokers: 0.
Created topic test3.
real	0m25.180s

and the resulting topic has the expected layout:

Topic: test3	PartitionCount: 3	ReplicationFactor: 1

Checklist

• I have explained the need for this PR and the problem it solves
• I have explained the changes or the new features added to this PR
• I have added tests corresponding to this change
• I have updated the documentation to reflect this change
• I have verified that this change is backward compatible (If not, please discuss on the APISIX mailing list first)