In one environment, with a constant load of 20k records for 10 mins, we see dropped records in 4.1.0 but not in 4.0.1. In other environments, we are not able to reproduce this. Is anyone else seeing this. Any thoughts on what the cause might be? fluent-bit.service

[Unit]
Description=Fluent Bit
Documentation=<https://docs.fluentbit.io/manual/>
Requires=network.target
After=network.target
# Disable the check for maximum allowed in an interval
StartLimitIntervalSec=0

[Service]
Type=simple
LimitNOFILE=13107200
ExecStart=/opt/fluent-bit/bin/fluent-bit -c /etc/fluent-bit/fluent_bit.yaml
MemoryMax=2G
# Allow for infinite restarts to support Fleet Management configuration changes fixing invalid configuration
Restart=on-failure
# Wait 30 second between restarts rather than 100ms
RestartSec=30s

[Install]
WantedBy=multi-user.target



fluent_bit.yaml

includes:
 - inputs_*.yaml
 - outputs_*.yaml
 - multiline_parsers_*.yaml
 - parsers_*.yaml

service:
 log_level: info
 http_server: on
 http_listen: 0.0.0.0
 http_port: 2021
 hot_reload: on
 log_file: /opt/fluent-bit/logs/fluent-bit.log
 storage.path: /opt/fluent-bit/flb-storage/
 storage.metrics: on
 storage.backlog.mem_limit: 300M
 storage.delete_irrecoverable_chunks: on

What's your config? You seem to be including things but no idea what they are. Why are you getting dropped records? It should tell you which will help narrow it down. What's your target? OS/container? You can also try a more recent version, e.g. 4.2.0. Performance will change between major versions but usually in a positive way!