Lasso - Lighthouse as a service

An API service built on top of lighthouse to automate running Lighthouse tests on large number of URLs in parallel. Utilizes Cloud Run and Cloud Tasks to distribute and run multiple tests across multiple containers, outputs results to a BigQuery dataset.

Features

✅ Bulk test 100s of URLs with lighthouse asynchronously

✅ Writes test results to a date partitioned BigQuery table

✅ Specify which resource requests to block from tests e.g. For running tests excluding 3rd party scripts or libraries

Setup

Deploying to Cloud Run

Create a build from the Docker file and create a new tag

docker build -t lasso .

Tag the built image for GCR targeting the cloud project

docker tag lasso gcr.io/[CLOUD-PROJECT-ID]/lasso:[TAG]

Push to GCR (Make sure the GCR API is enabled already)

docker push gcr.io/[CLOUD-PROJECT-ID]/lasso:[TAG]

To learn more about configuring a Cloud Run service, you can follow this guide.

The following ENV variables will need to be configured on cloud run:

Running Locally via docker compose

Setting Environment Variables

Set the path for the Google Cloud credentials of the project as an ENV variable in your system on GCP_KEY_PATH. This value will be picked up by the docker compose config mapped to a volume and set on the running container. The following ENV variables will need to be configured in docker-compose.yml:

Run

For local development, you can choose to run the project via docker compose. Running docker-compose up launches the service.

Running the image directly locally or on a server

PORT=8080 && docker run \
   -p 8080:${PORT} \
   -e PORT=${PORT} \
   -e BQ_DATASET=lh_results \
   -e BQ_TABLE=psmetrics \
   -e GOOGLE_APPLICATION_CREDENTIALS=/tmp/keys/lighthouse-service-09c62b8cd84e.json \
   -v $GOOGLE_APPLICATION_CREDENTIALS:/tmp/keys/lighthouse-service-09c62b8cd84e.json:ro \
   gcr.io/lighthouse-service/pagespeed-metrics

Using the service API

/audit (POST)

Runs one or more audits sequentially, utilizing a shared puppeteer instance between tests. Logs results to the configured BQ dataset table.

Parameters

Example

curl -X POST \
  http://127.0.0.1:8080/audit \
  -H 'content-type: application/json' \
  -d '{
  "urls": [
    "https://www.exampleurl1.com",
    "https://www.exampleurl1.com",
    ...
    ],
  "blockedRequests": [
    "https://www.someblockedrequestdomain.com"
    ]
  }'

/bulk-schedule (POST)

Schedules one or more audits to run asynchronously, utilizing Cloud Tasks. Each dispatched task calls /audit as a target to run and log the test.

Example

curl -X POST \
  http://127.0.0.1:8080/bulk-schedule \
  -H 'content-type: application/json' \
  -d '{
  "urls": [
    "https://www.exampleurl1.com",
    "https://www.exampleurl1.com",
    ...
    ],
  "blockedRequests": [
    "https://www.someblockedrequestdomain.com"
    ]
  }'

/active-tasks (GET)

Lists all the active audit tasks that are in queue along with the payload and status. Applies pagination to results.

Disclaimer

This is not an officially supported Google product.