Serve the Feed

The feed must be reachable over HTTPS at the well-known path on the apex domain:

https://example.com/.well-known/ojobpub.json

Redirects are allowed, including to another host or path, as long as the final response is the JSON document. Consumers (SourceTracker) follow at most 5 redirects.

There are three ways to serve the feed:

• Static file: the file sits in the web root.
• Redirect: the apex URL redirects to where the file is hosted.
• Reverse proxy: the web server proxies the path to an application or bucket.

Static file

Place the file in .well-known/ below the web root, e.g.:

/var/www/example.com/.well-known/ojobpub.json

Many setups redirect example.com to www.example.com and keep the path. In that case, placing the file in the www vhost is enough:

$ curl -sI https://example.com/.well-known/ojobpub.json
HTTP/2 301
location: https://www.example.com/.well-known/ojobpub.json

Redirect

Redirect the well-known path to wherever the feed is hosted:

server {
  server_name example.com www.example.com;

  location = /.well-known/ojobpub.json {
    return 302 https://jobs.example.com/feeds/ojobpub.json;
  }
}

<VirtualHost *:443>
  ServerName example.com
  ServerAlias www.example.com

  Redirect 302 "/.well-known/ojobpub.json" "https://jobs.example.com/feeds/ojobpub.json"
</VirtualHost>

example.com, www.example.com {
  redir /.well-known/ojobpub.json https://jobs.example.com/feeds/ojobpub.json 302
}

Use a temporary redirect (302/307) if the target may change.

Reverse proxy

A reverse proxy keeps the public URL stable and hides the origin. Use it when an application generates the feed, or when the file lives in object storage.

location = /.well-known/ojobpub.json {
  proxy_pass http://127.0.0.1:8080/api/ojobpub.json;
}

location = /.well-known/ojobpub.json {
  proxy_pass https://my-bucket.s3.eu-central-1.amazonaws.com/ojobpub.json;
  proxy_set_header Host my-bucket.s3.eu-central-1.amazonaws.com;
}

Response headers

Verify

curl -sSL -D- https://example.com/.well-known/ojobpub.json -o /dev/null   # status, redirects, headers
curl -sSL https://example.com/.well-known/ojobpub.json | python3 -m json.tool  # valid JSON

Then validate the downloaded document against the schema.