CircleCI
In CircleCI's Docker executor, the first image is where steps run. Every image after it is a service, and all of them share one network, so the tests reach every service on localhost.
How services work in CircleCI
- Hostname: always
localhost, with each service on its default port. Two services that use the same port can't share a job. - Readiness: CircleCI doesn't wait for services. Each job waits for the port with the
wait-forcommand below. CI=true: CircleCI sets it for every job.- Kafka: advertise
localhost:9092, since that's where the tests connect.
Complete .circleci/config.yml
All jobs are in one workflow with no requires, so they run in parallel. The --project names are the ones in Propmaster's playwright.config.ts; use yours.
yaml
version: 2.1
commands:
wait-for:
description: Wait until a TCP port accepts connections
parameters:
port: { type: integer }
steps:
- run:
name: Wait for port << parameters.port >>
command: |
for i in $(seq 1 90); do
node -e "require('net').connect(<< parameters.port >>, 'localhost').on('connect', () => process.exit(0)).on('error', () => process.exit(1))" && exit 0
sleep 2
done
echo "Port << parameters.port >> is not reachable"; exit 1
playwright:
description: Cached npm install, setup steps, Playwright tests, and results
parameters:
projects: { type: string }
npm-flags: { type: string, default: "" }
setup: { type: steps, default: [] }
steps:
- checkout
- restore_cache:
keys:
- npm-v1-{{ checksum "package-lock.json" }}
- npm-v1-
- run: npm ci --prefer-offline --no-audit --no-fund << parameters.npm-flags >>
- save_cache:
key: npm-v1-{{ checksum "package-lock.json" }}
paths: [~/.npm]
- steps: << parameters.setup >>
- run:
name: Playwright tests
command: npx playwright test << parameters.projects >> --reporter=line,junit
environment:
PLAYWRIGHT_JUNIT_OUTPUT_FILE: results/junit.xml
- store_test_results:
path: results
- store_artifacts:
path: playwright-report
jobs:
checks:
docker:
- image: node:24
steps:
- checkout
- restore_cache:
keys:
- npm-v1-{{ checksum "package-lock.json" }}
- npm-v1-
- run: npm ci --prefer-offline --no-audit --no-fund
- run: npm run typecheck
- run: npm run lint:tests
postgres:
docker:
- image: node:24
environment:
PW_DB_TYPE: postgresql
PW_DB_HOST: localhost
PW_DB_PORT: "5432"
PW_DB_NAME: testdb
PW_DB_USERNAME: testuser
PW_DB_PASSWORD: testpass
- image: postgres:15
environment:
POSTGRES_DB: testdb
POSTGRES_USER: testuser
POSTGRES_PASSWORD: testpass
steps:
- playwright:
projects: --project=integration-database
setup:
- wait-for: { port: 5432 }
- run: node scripts/create-test-schema.js # your schema setup
redis:
docker:
- image: node:24
environment:
PW_REDIS_HOST: localhost
PW_REDIS_PORT: "6379"
PW_REDIS_KEY_PREFIX: "test:ci-<< pipeline.number >>:" # ${VARS} aren't expanded here
- image: redis:7
steps:
- playwright:
projects: --project=integration-redis
setup:
- wait-for: { port: 6379 }
mongodb:
docker:
- image: node:24
environment:
PW_MONGODB_HOST: localhost
PW_MONGODB_DATABASE: testdb
- image: mongo:7
steps:
- playwright:
projects: --project=integration-mongodb
setup:
- wait-for: { port: 27017 }
kafka:
docker:
- image: node:24
environment:
PW_KAFKA_BROKERS: localhost:9092
- image: apache/kafka:latest
environment:
KAFKA_NODE_ID: 1
KAFKA_PROCESS_ROLES: broker,controller
KAFKA_LISTENERS: PLAINTEXT://0.0.0.0:9092,CONTROLLER://0.0.0.0:9093
KAFKA_ADVERTISED_LISTENERS: PLAINTEXT://localhost:9092
KAFKA_CONTROLLER_LISTENER_NAMES: CONTROLLER
KAFKA_LISTENER_SECURITY_PROTOCOL_MAP: PLAINTEXT:PLAINTEXT,CONTROLLER:PLAINTEXT
KAFKA_CONTROLLER_QUORUM_VOTERS: 1@localhost:9093
KAFKA_AUTO_CREATE_TOPICS_ENABLE: "true"
KAFKA_NUM_PARTITIONS: 1
KAFKA_OFFSETS_TOPIC_REPLICATION_FACTOR: 1
CLUSTER_ID: MkU3OEVBNTcwNTJENDM2Qk
steps:
- playwright:
projects: --project=integration-kafka
setup:
- wait-for: { port: 9092 }
browser:
docker:
- image: mcr.microsoft.com/playwright:v1.65.1-noble # must match @playwright/test
resource_class: large
steps:
- playwright:
# Skips native builds: only needed if the project installs better-sqlite3 (SQLite), which this image can't compile.
npm-flags: --ignore-scripts
projects: --project=browser-chromium --project=browser-firefox --project=browser-webkit
workflows:
tests:
jobs: [checks, postgres, redis, mongodb, kafka, browser]node:24 and the Playwright image both include git and ssh, which checkout needs. CircleCI's cimg/node images work too.
Other backends
Add a job like redis above with this service image:
| Backend | Service image and environment | Port to wait for | Variables on the first image |
|---|---|---|---|
| MySQL | mysql:8.0; MYSQL_DATABASE, MYSQL_USER, MYSQL_PASSWORD, MYSQL_ROOT_PASSWORD | 3306 | PW_DB_TYPE: mysql, PW_DB_HOST: localhost, PW_DB_PORT: "3306", name, user, password |
| MSSQL | mcr.microsoft.com/mssql/server:2022-latest; ACCEPT_EULA: "Y", MSSQL_SA_PASSWORD | 1433 (then sleep 10) | PW_DB_TYPE: mssql, PW_DB_HOST: localhost, PW_DB_PORT: "1433", PW_DB_NAME: master, PW_DB_USERNAME: sa, PW_DB_ENCRYPT: "false", PW_DB_TRUST_SERVER_CERTIFICATE: "true" |
| Neo4j | neo4j:5; NEO4J_AUTH: neo4j/testpassword | 7687 | PW_NEO4J_URI: bolt://localhost:7687, PW_NEO4J_USERNAME: neo4j, PW_NEO4J_PASSWORD: testpassword |
Secrets
Put real credentials in a context or in the project's environment variables. They reach every step as environment variables of the same name:
yaml
workflows:
tests:
jobs:
- staging-api:
context: propmaster-staging # holds PW_DB_PASSWORD, PW_GRAPHQL_AUTH_TOKEN, ...
filters: { branches: { only: main } }Tips
- Auto-cancel: turn on Project Settings → Advanced → Auto-cancel redundant workflows so a new push cancels the older run.
- Tests tab:
store_test_resultsadds the Tests tab and timing data. With timing data,circleci tests split --split-by=timingscan split a large suite acrossparallelismcontainers. - Memory: MSSQL needs about 2 GB. Use
resource_class: largefor its job. - Docs-only changes: CircleCI has no built-in path filter. Use the path-filtering orb if you need one.