A library designed for TDD while leveraging docker compose.
docker-compose-testkit helps writing tests by automating the creation and deletions of docker compose environments via test code.
Each test suite (or test!), may create it's own isolated compose network to test against, helping create stronger tests.
The library can work with any test runner, or just via code, the only requirement is docker compose v2 installed on the machine.
$ npm i --save-dev docker-compose-testkitOr alternatively
$ yarn add -D docker-compose-testkitGiven the following docker-compose.yml file
version: '3'
services:
nginx:
image: nginx:latest
ports:
- 80The following test will up the network before all the tests, and tear it down once the tests finish
import path from 'path'
import {fileURLToPath} from 'url'
import dockerCompose from 'docker-compose-testkit'
describe('testing using docker compose', () => {
const pathToCompose = path.join(
path.dirname(fileURLToPath(import.meta.url)),
'docker-compose.yml',
)
const compose = dockerCompose(pathToCompose)
beforeAll(compose.setup)
afterAll(compose.teardown)
it('should send a request to nginx', async () => {
// translates the internal address (i.e. nginx:80) to host's (e.g. 0.0.0.0:23424)
const nginxAddress = await compose.getAddressForService('nginx', 80)
expect((await fetch(`http://${nginxAddress}/`)).ok).to.be.true()
})
})Creates the compose object that will manage the network, once it is set up.
Specific options to set for the network and containers
servicesToStart: string[] - A list of cherry-picked services to start with the network (default: everything).
tailServices: string - Name of a service to start tailing its logs immediately, logs will be printed to stdout regardless of the original file descriptor.
projectName: string - Can be used to reconnect to an existing network from a previous run.
env: object - Key value object that will set the environment variables for the compose commands, by default containers will not have access to the host's env vars.
orphanCleanup: boolean - Wether or not to perform cleanup for containers that might have been left by a previous run (default true).
cleanup: boolean - Wether or not to perform a cleanup of the network after the test complete (disables the teardown) (default: true).
pullImages: boolean - Attempt to pull the images prior to docker compose up (default: false).
forceKill: boolean - Send a SIGKILL to all containers before running docker compose down, this can speed up the teardown significantly (default: true).
containerRetentionInMinutes: number - How long to keep this network from being picked up by the orphan cleanup, use alongside with cleanup: false to retain a network for debugging purposes (default: 5 minutes).
Returns Promise<Compose>.
A string representing the generated docker compose project name, pass it again to dockerCompose(pathToCompose, {projectName}) to reconnect to an existing network.
The path to the compose file as given to the dockerCompose(pathToCompose)
Runs cleanup for orphaned containers, networks and volumes. And then starts up the new network.
Returns Promise.
Tears down the current network.
Returns Promise.
Translates the internal address of a given service to the exposed one on the host's machine. For example nginx:80 to 0.0.0.0:52736.
The function will attempt to contact the service (similarly to Kubernetes liveness probe), to gauge wether the service is ready to accept connections.
Returns Promise<string>
healthCheck: undefined | false | string | (address, AbortError) => Promise<void> - A custom health check function that will be retried to gauge when the service is ready to be used.
undefinedis the default, will attempt to send a GET request to/.falsedisables the health check, the function will return with the address as normal.stringchanges the path of the GET request, can be used for/healthz.functiona custom function that defines the health check, it will be retried automatically, unless anAbortErroris sent. Functions arguments:- address: string - the addres of the service to test
- AbortError: (message: string) => AbortError - functions that creates an error that will stop the retries immediately
fiveHundedStatusIsOk: boolean - Wether to consider >500 status codes as healthy (default: false).
maxRetries: number - How many times to retry the health check, the retries have exponential backoff (default: 10).
Translates the inernal hostname (i.e. nginx) to the internal network IP, for example 172.17.0.1.
Returns Promise<string>.
Gets a list of the containers in the network.
Returns Promise<Container[]>.
Checks if a container exists for a specific service.
Returns Promise<boolean>.
Returns all the logs by a given service name, stdout and stderr will be interleaved.
Returns Promise<string>
Waits for a running service to finish and exit.
Returns Promise
anyExitCode: boolean - Wether to throw if the exit code is not 0 (default: false).
timeout: number - Total timeout to wait for in milliseconds (default: 5 minutes).
Runs a one-off command in a service (e.g. docker compose run).
commandWithArgs - An array of the command and arguments for it.
Returns Promise<ExecaReturnValue>
Starts a service in the network (e.g. docker compose start).
Returns Promise.
Stops a service in the network (e.g. docker compose stop).
Returns Promise.
Pauses a service in the network (e.g. docker compose pause).
Returns Promise.
Unpauses a service in the network (e.g. docker compose unpause).
Returns Promise.