fluentd-benchmark

Benchmark collection of Fluentd use cases.

For the simplest usecase, have a look at one_forward first.

Provided benchmarks

Testcase	Purpose
in_forward	Measure forwarding input rate
kafka-connect	Measure Kafka rate
keep_forward	Measure rate with keepalive
one_forward	Measure rate with single forwarder
out_forward	Measure forwarding output rate
out_kafka	Measure various output rates to Kafka
peer_deliver	Measure multi-receiver (separate) rate
round_robin_deliver	Measure multi-receiver (integrated) rate

docker-compose testing

For the usecases with Kafka or Keepalive a Docker Compose based test setup is provided, consisting of a YAML-based Docker Compose description and benchmark.sh wrapping it.

You would run it like this:

git clone https://github.com/fluent/fluentd-benchmark
cd fluentd-benchmark
./benchmark.sh one_forward

Please note that this does not use your current checked out version, but the master from https://github.com/fluent/fluentd-benchmark - adjust as needed.

This wrapper expects Docker Compose, GNU sed, GNU head and https://github.com/nferraz/st to be installed

(MacOS: brew install coreutils docker-compose gnu-sed st)