AnsweredAssumed Answered

Server specific limitations to requests to PI Web API

Question asked by dpfahler on Jan 8, 2020
Latest reply on Jan 17, 2020 by dpfahler

Hello Dear Community

I use the PI web API to retrieve data from a PI system.
I am sure that I am doing the sequence of requests in the correct order and for my use cases the queries work fine.
However, our customers are always encountering problems. These manifest themselves with empty responses or time-outs.
To get the data over the HTTP protocol we have to use some tricks:

  • Queries over larger time ranges are split up into several requests (potentially very many)
  • For that the number of data records are queried (GET/SUMMARY/summaryType=Count)
  • Multiple requests may be combined with batch requests
  • Or with "x-piwebapi-resource-address" the GET character limit is bypassed.
  • Parallel queries from multiple clients are sent to the system.

Since these tricks work well on our system, but not with the customers. I wonder what kind of server-side limitations you can think of that we don't respect.

  • Possible security mechanisms that get in our way?
  • Limits on parallel queries?
  • Time-outs for expensive requests?
  • Specific Request header/body lengths?

I am grateful for any help.