Flink1.18.0 with flink-cdc 3.1.0. Cause too many sleep connections in mysql.

[tatuke]

Hi everyone!
I found that deploying flink 1.18.0 in native k8s application mode, when deploying the application developed using datastream API and opening the task, it causes “too many connections” error for the source database and all other applications connecting to this database fail.
My database version is 5.7.32. and I am experimenting with a database that has almost no access. I found multiple SLEEP connections in the backend even when the source table is not modified or added. My database “max_connections=1000” . the application deployed by flink initiates about 70% of the connections. And most of them are sleep status.
These sleep connections last roughly 350 seconds or more. My database wait_timeout= 28800. so I wonder if I'm not getting this cdc related configuration right. The version of cdc I'm using is 3.1.0.
I checked issue #1015.So i look around the Apache jira.I can't find the corresponding version on jira.
Here are my deploy command:
./flink run-application
--target kubernetes-application
-Dkubernetes.cluster-id=dba-uat-old-orders
-Dkubernetes.container.image=flink-application18_order:1.3
-Dkubernetes.jobmanager.replicas=1
-Dtaskmanager.memory.process.size=4096m
-Dtaskmanager.numberOfTaskSlots=1
-Dkubernetes.namespace=flink-stream
-Dkubernetes.jobmanager.service-account=flink
-Dexternal-resource.limits.kubernetes.cpu=2000m
-Dexternal-resource.limits.kubernetes.memory=2Gi
-Dexternal-resource.requests.kubernetes.cpu=1000m
-Dexternal-resource.requests.kubernetes.memory=1Gi
-Dkubernetes.rest-service.exposed.type=NodePort
-c com.test.Order
local:///opt/flink/usrlib/flink-test-1.4-SNAPSHOT-jar-with-dependencies.jar

Will it be fixed in the next release?[FLINK-35524]
Any suggestions would be welcome!
Thank you!