ente/infra/workers

Cloudflare Workers

Source code for our Cloudflare Workers.

Workers are organized as Yarn workspaces sharing a common package.json and base tsconfig. They can however be deployed individually.

Deploying

Install dependencies with yarn.

If you have previously deployed, then you will have an old yarn.lock. In this case it is safe to delete and recreate using rm yarn.lock && yarn.

Then, to deploy an individual worker

• Login into wrangler (if needed) using yarn workspace health-check wrangler login

• Deploy! yarn workspace health-check wrangler deploy

Wrangler is the CLI provided by Cloudflare to manage workers. Apart from deploying, it also allows us to stream logs from running workers by using yarn workspace <worker-name> wrangler tail.

Creating a new worker

Copy paste an existing one. Unironically this is a good option because Cloudflare's template has a lot of unnecessary noise, but if really do want to create one from scratch, use npm create cloudflare@latest.

To import an existing worker from the Cloudflare dashboard, use

npm create cloudflare@2 existing-worker-name -- --type pre-existing --existing-script existing-worker-name

Logging

Attach the tail worker to your worker by adding

tail_consumers = [{ service = "tail" }]

in its wrangler.toml. Then any console.(log|warn|error) statements and uncaught exceptions in your worker will be logged to Grafana.