# Data loss in kafka broker

**URL:** <https://forum.confluent.io/t/data-loss-in-kafka-broker/8960>\
**Category:** Self-Managed Connectors\
**Created:** [30 October 2023 15:24 UTC](https://forum.confluent.io/t/data-loss-in-kafka-broker/8960 "2023-10-30T15:24:31Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![Uche](https://avatars.discourse-cdn.com/v4/letter/u/d26b3c/32.png) [@Uche](https://forum.confluent.io/u/Uche)\
**Post date:** [30 October 2023 15:24 UTC](https://forum.confluent.io/t/data-loss-in-kafka-broker/8960/1 "2023-10-30T15:24:32Z")

</div>

Please what configuration should I use to ensure that I don’t lose data when replicating?  
I have a source table that has 7 M rows but only about 6 M plus gets to the destination table. I think the consumer just idles while waiting. I have been trying to throttle the producer and consumer to produce and consume equal amount of data to reduce the idle time.  
My producer config:  
connector.class=io.debezium.connector.sqlserver.SqlServerConnector  
tasks.max=1  
include.schema.changes=true  
tombstones.on.delete=true  
topic.prefix=adc-  
decimal.handling.mode=double  
schema.history.internal.kafka.topic=schema-changes.adc  
include.schema.comments=false  
poll.interval.ms=10  
value.converter=io.confluent.connect.avro.AvroConverter  
key.converter=io.confluent.connect.avro.AvroConverter  
snapshot.lock.timeout.ms=5000  
database.encrypt=false  
database.user=${file:/etc/data/xxxxxxxx:username}  
database.names=xxxxxxx  
max.message.bytes=10485880  
acks=all  
time.precision.mode=connect  
schema.history.internal.kafka.bootstrap.servers=kafka:9092  
database.port=1433  
database.hostname=${file:/etc/data/xxxxxx:hostname}  
database.password=${file:/etc/data/xxxxxx:password}  
table.include.list=xxxxxx  
snapshot.mode=initial

And my Consumer config  
connector.class=io.confluent.connect.jdbc.JdbcSinkConnector  
connection.password=${file:/etc/data/xxxxxx:password}  
tasks.max=5  
decimal.handling.mode=double  
enable.auto.commit=true  
auto.evolve=true  
value.converter=io.confluent.connect.avro.AvroConverter  
session.timeout.ms=60000  
insert.mode=upsert  
key.converter=io.confluent.connect.avro.AvroConverter  
max.poll.records=10000  
topics=xxxxx  
heartbeat.interval.ms=3000  
delete.enabled=true  
connection.user=${file:/etc/data/xxxxxx:username}  
fetch.min.bytes=16384  
fetch.max.bytes=52428800  
auto.create=true  
connection.url=${file:/etc/data/xxxxxx:connectionurl}  
max.poll.interval.ms=100  
pk.mode=record\_key  
pk.fields=id

Please what else do I need to ensure that data is not lost?