locust-on-databricks 0.2.0

Run locust on single node or distributed on Databricks.

locust-on-databricks

Installation

pip install locust-on-databricks

Usage

OOTB Examples

For OOTB please look at examples in

1. 01 Locust Test Single Node.py
2. 02 Locust Test Distributed.py

The dummy example locustfile is here: locustfile.py

API Usage

Construct the runner:

from lod import LocustRunner

runner = LocustRunner(
    locustfile_path="locustfile.py",  # path to your locustfile
    # port=8089, # port is optional
)

Construct your initial swarm:

runner = runner.set_initial_swarm(
    host="https://google.com",
    user_count=10,
    spawn_rate=2,
    # run_time = "5m"
)

Start your locust server:

runner.start_locust()

Wait for running locust swarm to finish

runner.block_until_end_of_swarm(
    timeout_in_seconds=300,
    # check_every_n_seconds=5, # poll every n seconds default is 5s
)

Capture results/snapshots into a directory

This will create:

1. request metrics csv
2. failures csv
3. exceptions csv
4. html report

runner.save_artifacts(
    # save_directory="artifacts" # default is artifacts but recommended you save to volumes
    # file_prefix="" # default is empty prefix to your artifacts
)

Stop your locust server:

runner.stop_locust()

Search for "Access Locust Web UI at" in the output and open the link in your browser.

Distributed Runner

Construct a distributed runner:

runner = runner
.distributed()
.set_initial_swarm(
    host="https://google.com",
    user_count=10,
    spawn_rate=2,
    # run_time = "5m"
)

Construct a distributed runner with custom worker to core ratio:

runner = runner
.distributed(process_to_core_count_ratio=2.0)
.set_initial_swarm(
    host="https://google.com",
    user_count=10,
    spawn_rate=2,
    # run_time = "5m"
)

process_to_core_count_ratio is defaulted to 2 and it will spin up twice the number of workers as you have cores. You can tweak this as needed.

Run and stop new swarms

Run a new swarm:

runner.run_swarm(
    host="https://google.com",
    user_count=10,
    spawn_rate=2,
    # run_time = "5m"
)

Stop a running swarm

runner.stop_swarm()

Limitations

Starting a new swarm in the ui is bugged and the swarm rest api from the ui is not working.

Disclaimer

locust-on-databricks is not developed, endorsed not supported by Databricks. It is provided as-is; no warranty is derived from using this package. For more details, please refer to the license.