Deploy a Worker

An AVS operator is expected to serve proving requests from Lagrange network by running a Worker binary. The Lagrange sequence sends through the Gateway tasks to workers containing all the necessary inputs to generate zkproofs. The worker continuously listens for such tasks and generate proofs accordingly.

This is a step by step guide to deploy your own worker.

---

If you prefer video content you can watch this video tutorial

---

Pre-Requirements

There are 3 different types of workers. Every type can generate a specific type of proof. You can choose your infrastrcture depending on the type of proof you would like to be able to generate.

Worker Type	Cores (Dedicated)	vCPU (Cloud)	Memory	Disk	Internet	Proposed Hetzner instance
small	8	20	40 GB	60GB	✅	AX-52
medium	8	40	80 GB	60GB	✅	AX-52
large	16	90	180 GB	60GB	✅	AX-102

Setup Steps

1. Install Docker by following this guide
2. Clone this repo

git clone https://github.com/Lagrange-Labs/worker.git
cd worker

3. Generate a Lagrange key

#Generate the key
docker compose run --rm worker avs new-key

You should see something like this

Input password for Lagrange key:
new Lagrange keystore stored under config/lagr_keystore.json
	public_key = "66779122cb188d1e70889278ffb3ee8ee024d4a23742a59afbdaaa096fc5135c3a14d897de87d1c56adbe029619c231416e7dfb3f1de0a542dd8ac7f4748ce07"

4. Add the Ethereum keystore under config/priv_key.json. * This is the operator keystore generated by EigenLayer cli
5. Register the Lagrange key on-chain by signing it with the AVS operator key

docker compose run --rm worker avs register

6. You are now registered, you can go ahead and remove the keystore from config/priv_key.json
7. Add the Lagrange password to the .env file so that the service can be restarted without any manual intervention.

# set the AVS__LAGR_PWD key with the password you have set
vim .env

Run the worker Steps

1. Run the worker

docker compose up -d

Observability

Metrics

The worker exposes the prometheus metrics by default on port 9000

Dashboard

Starting from worker version v0.2.1, you can import this grafana dashboard