Exponential backoff for SPIKE Keepers' contribution endpoints

[v0lkan]

This is how SPIKE keepers send their contribution

      _, err = net.Post(client, contributeUrl, md)
      for err != nil {
      time.Sleep(5 * time.Second)
      _, err = net.Post(client, contributeUrl, md)
      if err != nil {
      log.Log().Info("contribute",
      "msg", "Error sending contribution. Will retry",
      "err", err,
      )
      time.Sleep(5 * time.Second)
      }
      }

while this retry logic works, it is better to exponentially back off so that during a network partition we can give time for the system to breathe and recover.

There are examples of exponential backoff in the source code.

The point of this issue is to align with those examples, rather than implementing a backoff algorithm from scratch.