playwright-remote-orchestrator

Purpose

• This was built as a method to run multiple tests against any number of servers. It will pick a random server and port and run the checks against them. It was built as playwright currently has a limitation where one can only run checks on the same box as where the code lives. This allows the checks to be put into docker containers and run centeralized while the actual orchestrators are ran in a number of places.

Limitations

• Currently it does not know if the remote nodes are up or down. Future state it will use the failure counters + a pinger to mark them up / down.
• queue.db needs to be shared along with the src files between the different containers. this is because it is using sqlite databases to determine which ports and hosts are free.

Building

1. Run docker-compose build

Running the checks with local server

1. uncomment the CMD line in docker/python/Dockerfile
2. Build the image and it should start if you do docker-compose up

• Note this has not been tested

Running a remote server

1. Install playwright server on the remote executioner host. Can be found here
2. once installed, start the server by running node_modules/.bin/playwright.cmd run-server --port <port number> --reuse-browser false

Configuration

1. modify src/database.yaml with the servername, start port, and end port
2. append --fr true to the docker-compose file and run it once. afterwords remove the --fr true. This is used to initialize the database the orchestrator uses to determine if ports are open or not.

Creating a check

• look at full/example/example.py for reference. If you need the logs, first make the log directory then add it to the docker-compose volumes config - ./logs/example_playwright/:/app/logs/.
• This check scrapes docusign and looks to see the latest status on the Docusign Nodes.