#885 Watch Sub / UnSub

James Tice Thu 11 Feb 2021

After having spent a good amount of time working with watches, a few questions still seem to remain.

  1. In a polling scenario with a large volume of points, is it generally more efficient to create a large number of watches with a small number of points each, or a small number of watches with a large number of points each? Example scenarios could be putting all points on 1 watch or putting each device's points on its own watch.
  2. Is there any way to request more data to come out of watch poll operations besides just id, curVal, and curStatus? We are currently using reads instead of watch polls (with the exception of an occasional watch poll to keep the sub active) to get the data we need. If we could request axStatus and other tags we care about during watch polls, it would dramatically simplify this polling process.
  3. Is there a way to request a list of currently active watches?

Also, thanks for putting together a forum where we can ask these questions!

Brian Frank Thu 11 Feb 2021

James, I think these questions are asked in the context of a specific implementation correct? Is this specific to nhaystack's implementation?

James Tice Fri 12 Feb 2021

Correct, we are currently only working with nhaystack if that helps at all

Richard McElhinney Thu 11 Mar 2021

Hi James,

I would actually suggest that for a very large volume of points that polling using a REST API may not be the optimal solution.

We have been working with the SparkplugB standard and are using it to send standardised messages over MQTT, which can contain Haystack tags and metadata, for multiple sites.

Let me know if you're interested and I can reach out offline.

Login or Signup to reply.