Forum Discussion

mvitale's avatar
mvitale
New Contributor
6 years ago

Salesforce Subscriber Snap Issue

Hello,

We’re using the Salesforce Subscriber Snap to subscribe to a Platform Event in our Org and have noticed that it will suddenly stop picking up new events after a period of time (typically after the pipeline has been running for a few weeks). Stopping and re-starting the pipeline results in things working again.

We do have an error pipeline defined, so I do not believe things are failing due to an error. It just seems to stop “listening” for new events.

Has anyone experienced this or have any recommendations?

Thanks!

9 Replies

  • I can confirm that we have a new fix that is currently in testing. You can expect a new patch release soon. QA for this snap takes time given the nature of the issues. Thanks for your patience.

    • lindsay's avatar
      lindsay
      New Contributor

      By any chance has similar behavior been reported for the publisher snap for platform events? We are investigating an issue with delayed publishing of platform events where no errors were reported.

      • ptaylor's avatar
        ptaylor
        Employee

        Delayed by what amount of time? Should there be errors reported if it’s just delayed? What version of the snap are you using?

  • nravon's avatar
    nravon
    New Contributor

    It’s a bug in the Snap - we have experienced the same issue since your time and we have worked with Snap support to get this resolved and the good news is - it should be fixed June 16th.

    The issue is with the way the underlying frameworks re-negotiate a handshake with salesforce via CometD (and / or via the Salesforce EMP framework) - basically after a time when the token expires - SnapLogic will try and renegotiate but it simply will not reconnect without a retart of the pipeline. You do not get an error in the pipeline but rather a warning and it will simply hang there never reconnecting.

    • nravon's avatar
      nravon
      New Contributor

      NOTE: there is a fix in the wild for this now. Speak to support and they should be able to apply it to your instance(s) for testing.

  • NAl's avatar
    NAl
    Contributor

    Not sure if this fix is in place - can SnapLogic support confirm?
    We’ve just experienced this issue and had to restart the pipeline.