Forum Discussion

pleblond's avatar
pleblond
New Contributor
5 years ago

Currently experiencing large load times

we currently have a pipe that queries salesforce via SOSL for search results. Using a normal *non ultra) pipeline task with a bearer token. When are using json to recieve and parse information then returning (non mapped) json to the requester.

This whole process takes 10 seconds which in production is an eternity.
Sample request:
{
“content”:
{
“language”:“en_US”,
“region”:“North_America__c”,
“customer-type”:“Customer__c”,
“keyword”:“the”,
“requiredFields”:“test1,test2”
}
}

Sample result:

[
{
“Title”: “test1”,
“ID__c”: 84
},
{
“Title”: “test2”,
“ID__c”: 75
},
{
“Title”: “test3”,
“ID__c”: 132
}
]

The results are very small so is the packet being sent in. Its also worth noting the pipe has a router determining a URL parameter for path that has 9 options.

2 Replies

  • tlikarish's avatar
    tlikarish
    Employee
    5 years ago

    Can you break down how you’re measuring the 10 seconds?

    Mind sharing a breakdown of the pipeline execution according to the Dashboard? If you don’t want to share publicly, you can reach out to support and they can share the results confidentially with us.

    If you click on the pipeline status in the dashboard

    you’ll get a break down of the pipeline’s execution, which may give some hints at where the pipeline is spending time.

    A general rule of thumb:

    1. If there is a big blue bar, then one of the snaps is taking a long time to process documents.
    2. If there are big green bars, snaps are waiting for input and something is taking a long time to load documents.
    3. If there are big purple bars, then one of the snaps is taking a long time to write out documents.

  • mbowen's avatar
    mbowen
    Employee
    5 years ago

    Hi Paul:

    I would suggest logging a support ticket for this. These kind of issues often require some research. Being able to inspect pipeline execution and logs super helpful. 10 seconds does seem less responsive for this query, but again, not familiar with size of data, etc. Actually, the pipeline execution would help identify where most of the time was spent.