I have a GeoEvent set up as follows.
Poll for a new feature (every 10 seconds, on hosted feature service through Survey123) > Using the imported service definition from that HFS > GeoEvent Service > Send Email (based on new feature, formatted in HTML).
Method to Identify Incremental Updates: ObjectID
Portal Update Interval: 1 Min
This IS set up on a Cluster (we are in the process of migrating our GeoEvent to one machine).
There were 8 e-mails that were sent out using the service throughout this week (no submissions before this week). They were all sent at appropriate times based on when the Survey123 Forms were submitted. (nothing to be surprised about here, everything seems to be working fine)
Now 2/23, all 8 emails (duplicates) were resent with no particular pattern.
2 at 12:20pm (dates 2/22 & 2/21)
5 at 12:25pm (dates 2/21 & 2/21 & 2/22 & 2/22 & 2/22)
1 at 12:35pm (date 2/21)
There are still only 8 features in my Survey123 HFS, none have been updated.
It is strange they are sending in increments of 5 minutes.
Is this likely an issue with the cluster? From what I can tell there were 8 inputs on PRD1 originally, then PRD2 finally caught up (in increments of 5 minutes) and sent those 8 out again. It appears as if PRD1 is not informing PRD2 that ObjectIDs 1-8 were processed. Then, PRD2 kicks in, thinks they were NOT processed, and fires them out. That does not explain the 5 min increments but might point to issues with clustering.