• State Not Answered
  • Replies 1 reply
  • Subscribers 19355 subscribers
  • Views 87 views
  • Users 0 members are here
Options
Related

Impact of large PI Web API or DataLink on Data Archive performance

leobar95

Hi all,

I’m trying to better understand how the PI Data Archive handles heavy client requests, especially from PI DataLink and PI Web API.

In some cases, users generate large queries — long time ranges, dense sampling, or many tags at once — and I’m concerned about the potential impact on archive performance (CPU, memory, or buffering delays on interfaces).

I’m aware of parameters like ArcMaxCollect and Archive_MaxQueryExecutionSec, but I’d like to know:

Are there additional mechanisms that prevent DataLink or Web API queries from overloading the Data Archive?

How does the system behave when several large client requests run in parallel?

Appreciate any technical insights or references to official documentation.

Thanks,
Leonardo

p.s. at the end of the day, from the Data Archive perspective, is a query from either DataLink or the WebAPI identical?

Parents
  • 0

    There are tunning parameters in place to prevent a client for overwhelming the PI Data Archive and using all of the archive threads.  You can read about PIArchss_MaxThreadsPerClietnQuery here:

    How to limit expensive queries to the PI Server archive

    If you have many users, you might want to increase the PIarchss_threadCount.  It should be increased to no more than the number of processors available.  I believe the default is eight.  There are other parameter, such as the "Archive_BulkQueryThreadNumber", which limits the number of threads available for bulk request.  Typically, the default is fine. 

    My understanding is that a queries from DataLink or the WebAPI are pretty identical.

Reply
  • 0

    There are tunning parameters in place to prevent a client for overwhelming the PI Data Archive and using all of the archive threads.  You can read about PIArchss_MaxThreadsPerClietnQuery here:

    How to limit expensive queries to the PI Server archive

    If you have many users, you might want to increase the PIarchss_threadCount.  It should be increased to no more than the number of processors available.  I believe the default is eight.  There are other parameter, such as the "Archive_BulkQueryThreadNumber", which limits the number of threads available for bulk request.  Typically, the default is fine. 

    My understanding is that a queries from DataLink or the WebAPI are pretty identical.

Children
No Data