I have a REST API that implements server-side pagination returning the following data:

{
  items: [
    {
      id: 1,
      name: 1
    },
    {
      id: 2,
      name: 2
    }
  ],
  nextToken: <some-hash-key> // cursor-based
}

How should the client application refresh the list if the resource gets updated (client isn't aware of the update, so this is a pull model)? I have some ideas:

1. fetch all resources at a regular interval (like every 10 seconds)
2. maintain a session ID that rotates every N minutes. When a new session is created, fetch all resources

Both approaches are fundamentally the same idea. The first approach is more costly but allows more real-time updates. The second approach is based on session ID which I think is more idiomatic, but no real-time updates. Is there any other approach?