I had filebeat running and is configured with kafka as ouput. Everything is working fine but when connection with kafka server is broke, filebeat stores the info of last log line read in its registry. But when connection is resumed, filebeat does not ingest old data.
Below is my configuration of filebeat:

filebeat.yml

filebeat.inputs:
- input_type: log
  close_inactive: 120m
  paths: [/home/test/files.log]
  json.keys_under_root: true
  enabled: true
  fields:
   topic: test

output.kafka:
   hosts: ["192.168.0.155:9092"]
   topic: "%{[fields.topic]}"
   required_acks: 1
   bulk_max_size: 100000
   broker_timeout: 84600
   timeout: 84600
   retry.max: 10000
   retry.backoff: 84600
   channel_buffer_size: 100000

Filebeat version is 7.16.3
Can anyone tell how to avoid this data loss?