GitHub repository: tinkerbell/hegel.

Hegel is Tinkerbell’s metadata store, supporting storage and retrieval of metadata over gRPC and HTTP. It also provides a compatible layer with the AWS EC2 metadata format.

Metadata is user-defined as part of the hardware data that makes up a workflow.

Using Hegel

You can access Hegel in a Tinkerbell setup at the Provisioner’s IP address /metadata. You can use cURL to retrieve the metadata it stores.

Hegel by default exposes an HTTP API on port 50061. You can interact with it via cURL

curl <hegel_ip>:50061/metadata

You can also retrieve a AWS EC2 compatible format uses from /meta-data.

$ curl <hegel_ip>:50061/<date>/meta-data

For example, if you are using the Vagrant Setup, Hegel runs as part of the Provisioner virtual machine with the IP: When the Worker starts and if you have logged in to osie using the password root you can access the metadata for your server via cURL:

curl -s | jq .
    "facility": {
        "facility_code": "onprem"
    "instance": {},
    "state": ""

Or in AWS EC2 format:

curl -s

If you look at the hardware-data.json that we used during the Vagrant setup you will find the facility_code=onprem as well.

Other Resources

Every cloud provider is capable of exposing metadata to servers that you can query as part of your automation, usually via HTTP. Some examples: