progress update's performance issue

UI will initiate an individual progress update query to the backend for every installed machine every 2 maybe 3 seconds, which will pose performance issue both for the backend and browser if we are talking about hundreds or thousands of nodes.
Suggest aggregate the queries in one request also change api accordingly.
We also notice that if progress update query is send immediately after deploy is initiated, there will be a probability that database is locked because there are many database operations going on at the early phase of deployment. Maybe better off delay the query after deployment is started and increase the interval of queries.