Hey @Anatoliy, welcome to elastic discuss
unfortunately there's not a deterministic way of gathering logs from old, not running containers.
auto discover feature work only with running and newly started containers
If you use docker input and specify a specific container id you might be able to ingest logs. But just in case docker did not get rid of the metadata
you can configure docker input as stated here
filebeat.inputs:
- type: docker
containers.ids:
- '8b6fe7dc9e067b58476dc57d6986dd96d7100430c5de3b109a99cd56ac655347'
Also please make sure to format any configuration you post here. Theres a button </> to do that. Sometimes the issue is with indentation of the configuration which is hard to spot if it is not correctly formatted