Hey, I'd like to ask you if you have experience with error

could not enqueue records into the ring buffer

(see https://fossies.org/linux/fluent-bit/src/flb_input_chunk.c line 1922). We are getting this error frequently along with the less frequent

broken connection to <aws_s3_bucket_url>

. We experience these errors on both small and big AWS EKS clusters (fluentbit running as daemonset) Our outputs configuration looks like this:

upload_timeout           5m

total_file_size          50M

auto_retry_requests      true

retry_limit              5

store_dir                <path_to_store_dir>

use_put_object           On

compression              gzip

store_dir_limit_size     512M

We tried several recommendations from the Fluentbit documentation along with upgrading to the latest stable version. Are we hitting limits of AWS S3 bucket or we need to change our buffering configuration? According to our metrics we are frequently dropping records. Any help or directions would be appreciated.