cometbft-load-test

cometbft-load-test is a framework for load testing CometBFT-based networks, forked from tm-load-test.

The structure/format of transactions sent to a CometBFT-based network are specific to the ABCI application running on that network, so while cometbft-load-test comes with built-in support for the kvstore ABCI application (as an example), you have to build your own clients for your own apps.

Requirements

cometbft-load-test is currently tested using Go v1.22 and CometBFT v0.38.

Usage

Step 1: Create your project

You have to create your own load testing tool for your own ABCI application by importing the cometbft-load-test package into a new project.

mkdir -p /your/project/
cd /your/project
go mod init github.com/you/my-load-tester

Step 2: Create your transaction generator

Create a client that generates transactions for your ABCI app. For an example, you can look at the kvstore client code. Put this in ./pkg/myabciapp/client.go

package myabciapp

import "github.com/cometbft/cometbft-load-test/pkg/loadtest"

// MyABCIAppClientFactory creates instances of MyABCIAppClient
type MyABCIAppClientFactory struct {}

// MyABCIAppClientFactory implements loadtest.ClientFactory
var _ loadtest.ClientFactory = (*MyABCIAppClientFactory)(nil)

// MyABCIAppClient is responsible for generating transactions. Only one client
// will be created per connection to the remote CometBFT RPC endpoint, and
// each client will be responsible for maintaining its own state in a
// thread-safe manner.
type MyABCIAppClient struct {}

// MyABCIAppClient implements loadtest.Client
var _ loadtest.Client = (*MyABCIAppClient)(nil)

func (f *MyABCIAppClientFactory) ValidateConfig(cfg loadtest.Config) error {
    // Do any checks here that you need to ensure that the load test
    // configuration is compatible with your client.
    return nil
}

func (f *MyABCIAppClientFactory) NewClient(cfg loadtest.Config) (loadtest.Client, error) {
    return &MyABCIAppClient{}, nil
}

// GenerateTx must return the raw bytes that make up the transaction for your
// ABCI app. The conversion to base64 will automatically be handled by the
// loadtest package, so don't worry about that. Only return an error here if you
// want to completely fail the entire load test operation.
func (c *MyABCIAppClient) GenerateTx() ([]byte, error) {
    return []byte("this is my transaction"), nil
}

Step 3: Create your CLI

Create your own CLI in ./cmd/my-load-tester/main.go:

package main

import (
    "github.com/cometbft/cometbft-load-test/pkg/loadtest"
    "github.com/you/my-load-tester/pkg/myabciapp"
)

func main() {
    if err := loadtest.RegisterClientFactory("my-abci-app-name", &myabciapp.MyABCIAppClientFactory{}); err != nil {
        panic(err)
    }
    // The loadtest.Run method will handle CLI argument parsing, errors,
    // configuration, instantiating the load test and/or coordinator/worker
    // operations, etc. All it needs is to know which client factory to use for
    // its load testing.
    loadtest.Run(&loadtest.CLIConfig{
        AppName:              "my-load-tester",
        AppShortDesc:         "Load testing application for My ABCI App",
        AppLongDesc:          "Some long description on how to use the tool",
        DefaultClientFactory: "my-abci-app-name",
    })
}

For an example of very simple integration testing, you could do something similar to what's covered in integration_test.go.

Step 4: Build your CLI

Then build the executable:

go build -o ./build/my-load-tester ./cmd/my-load-tester/main.go

Running your load testing tool

A cometbft-load-test-based load testing application can be executed in one of two modes: standalone, or coordinator/worker.

NB: In all of the following examples, replace cometbft-load-test with the name of your load testing application you have built (e.g. my-load-tester).

Standalone Mode

In standalone mode, cometbft-load-test simply broadcasts transactions to a single endpoint from a single binary:

cometbft-load-test -c 1 -T 10 -r 1000 -s 250 \
    --broadcast-tx-method async \
    --endpoints ws://cmt-endpoint1.somewhere.com:26657/websocket,ws://cmt-endpoint2.somewhere.com:26657/websocket

To see a description of what all of the parameters mean, simply run:

cometbft-load-test --help

Coordinator/Worker Mode

In coordinator/worker mode, which is best used for large-scale, distributed load testing, cometbft-load-test allows you to have multiple worker machines connect to a single coordinator to obtain their configuration and coordinate their operation.

The coordinator acts as a simple WebSockets host, and the workers are WebSockets clients.

On the coordinator machine:

# Run cometbft-load-test with similar parameters to the standalone mode, but now
# specifying the number of workers to expect (--expect-workers) and the host:port
# to which to bind (--bind) and listen for incoming worker requests.
cometbft-load-test \
    coordinator \
    --expect-workers 2 \
    --bind localhost:26670 \
    -c 1 -T 10 -r 1000 -s 250 \
    --broadcast-tx-method async \
    --endpoints ws://cmt-endpoint1.somewhere.com:26657/websocket,ws://cmt-endpoint2.somewhere.com:26657/websocket

On each worker machine:

# Just tell the worker where to find the coordinator - it will figure out the rest.
cometbft-load-test worker --coordinator localhost:26680

For more help, see the command line parameters' descriptions:

cometbft-load-test coordinator --help
cometbft-load-test worker --help

Endpoint Selection Strategies

An endpoint selection strategy can now be given to cometbft-load-test as a parameter (--endpoint-select-method) to control the way in which endpoints are selected for load testing. There are several options:

1. supplied (the default) - only use the supplied endpoints (via the --endpoints parameter) to submit transactions.
2. discovered - only use endpoints discovered through the supplied endpoints (by way of crawling the CometBFT peers' network info), but do not use any of the supplied endpoints.
3. any - use both the supplied and discovered endpoints to perform load testing.

NOTE: These selection strategies only apply if, and only if, the --expect-peers parameter is supplied and is non-zero. The default behaviour if --expect-peers is not supplied is effectively the supplied endpoint selection strategy.

Minimum Peer Connectivity

cometbft-load-test can wait for a minimum level of P2P connectivity before starting the load testing. By using the --min-peer-connectivity command line switch, along with --expect-peers, one can restrict this.

What this does under the hood is that it checks how many peers are in each queried peer's address book, and for all reachable peers it checks what the minimum address book size is. Once the minimum address book size reaches the configured value, the load testing can begin.

Monitoring

cometbft-load-test exposes a number of Prometheus metrics when in coordinator/worker mode, but only from the coordinator's web server at the /metrics endpoint. So if you bind your coordinator node to localhost:26670, you should be able to get these metrics from:

curl http://localhost:26670/metrics

The following kinds of metrics are made available here:

• Total number of transactions recorded from the coordinator's perspective (across all workers)
• Total number of transactions sent by each worker
• The status of the coordinator node, which is a gauge that indicates one of the following codes:
  • 0 = Coordinator starting
  • 1 = Coordinator waiting for all peers to connect
  • 2 = Coordinator waiting for all workers to connect
  • 3 = Load test underway
  • 4 = Coordinator and/or one or more worker(s) failed
  • 5 = All workers completed load testing successfully
• The status of each worker node, which is also a gauge that indicates one of the following codes:
  • 0 = Worker connected
  • 1 = Worker accepted
  • 2 = Worker rejected
  • 3 = Load testing underway
  • 4 = Worker failed
  • 5 = Worker completed load testing successfully
• Standard Prometheus-provided metrics about the garbage collector in cometbft-load-test
• The ID of the load test currently underway (defaults to 0), set by way of the --load-test-id flag on the coordinator

Aggregate Statistics

You can write simple aggregate statistics to a CSV file once testing completes by specifying the --stats-output flag:

# In standalone mode
cometbft-load-test -c 1 -T 10 -r 1000 -s 250 \
    --broadcast-tx-method async \
    --endpoints ws://cmt-endpoint1.somewhere.com:26657/websocket,ws://cmt-endpoint2.somewhere.com:26657/websocket \
    --stats-output /path/to/save/stats.csv

# From the coordinator in coordinator/worker mode
cometbft-load-test \
    coordinator \
    --expect-workers 2 \
    --bind localhost:26670 \
    -c 1 -T 10 -r 1000 -s 250 \
    --broadcast-tx-method async \
    --endpoints ws://cmt-endpoint1.somewhere.com:26657/websocket,ws://cmt-endpoint2.somewhere.com:26657/websocket \
    --stats-output /path/to/save/stats.csv

The output CSV file has the following format at present:

Parameter,Value,Units
total_time,10.002,seconds
total_txs,9000,count
avg_tx_rate,899.818398,transactions per second

Development

To run the linter and the tests:

make lint
make test

Integration Testing

Integration testing requires Docker to be installed locally.

make integration-test

This integration test:

1. Sets up a 4-validator, fully connected CometBFT-based network on a 192.168.0.0/16 subnet (the same kind of testnet as the CometBFT localnet).
2. Executes integration tests against the network in series (it's important that integration tests be executed in series so as to not overlap with one another).
3. Tears down the 4-validator network, reporting code coverage.

License

Copyright 2023 CometBFT team and contributors

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.