I'm facing many parasite update subscription API events for Key_Feature__c object.

Some of them are happening on the records that describe the activities finished in deep past. These records do not change, but still generate the update events. I know some ids of them.

My hypothesis is that some automation calls update keyFeature without changing any fields.

It might be anything - a flow or an Apex class. Unfortunately the org has hundreds of places where we save the Key_Feature__c. And lots of valid processes currently running.

How can I identify the automation pieces those are performing unnecessary saves?

What kind of tracking may I turn on or implement by myself?

  • 3
    Check the Modified By details - this will at least confirm that the records are being updated and show the contextual user against which the "no-op" update is being performed. Other than turning on debug logging and looking for database update messages and apex/flow execution tracing I can't think of a way to find where this is happening. – Phil W Apr 24 at 7:42

Is this occurring in a sandbox as well as production? It would be a lot easier to take additional steps in a sandbox to detect the unwanted updates. These could also be applied in production, but are more likely to have unwanted side effects.

You say that Key_Feature__c records shouldn't change. Especially the older ones. I'd use that as a basis for finding the problem processes.

In a sandbox you could create automation to prevent updates to older Key_Feature__c records. This might be an Apex trigger, a flow, a validation rule, etc... You want to check how old the record being updated is. If it is past the threshold, say a day old, then throw and error or otherwise prevent the transaction from completing.

E.g. with a trigger

trigger PreventUpdatesToOldKeyFeatures on Key_Feature__c (before update) {

    for(Key_Feature__c kf : Trigger.new) {
        if(kf.CreatedDate.addDays(1) > Date.today()) {
            kf.addError('Attempt to update old Key Feature');


Of course, a trigger may seem like overkill here. A Validation rule could be equally effective.

You might also want to check if there are any updates occurring to records with a master-detail relationship to Key Features.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.