Hello everyone. I am an admin troubleshooting some SSO issues, but I am also working on a small script to pull our daily interaction counts via the Analytics API. I am seeing a discrepancy where some conversations that started just before midnight are being counted twice, or are missing entirely from my daily reports. I suspect I am not handling the ‘Interval’ boundaries correctly in my query. Should I be using UTC or my local time zone for the interval string, and how do I handle conversations that cross the midnight boundary?
Hello Dia57! I am a routing optimization engineer and I have seen this interval problem break our skills-based reports. The Analytics API always uses UTC for the interval boundaries. If you want to align with your local time, you must calculate the UTC equivalent for your midnight. For conversations that cross the boundary, the analytics engine will include the conversation in whichever interval the interaction ended in, unless you use a more specific filter. I always recommend using a small ‘Overlap’ in your queries to ensure you catch everything, and then deduplicate the interaction IDs in your script.
Greetings. I build flows for many clients and I have spent a lot of time with the analytics data. To follow up on Sor88, you should also look at the conversationStart and conversationEnd fields in the response. If you want to be one hundred percent accurate for your daily reports, you should query a twenty-six hour window and then filter the results in your script based on the exact conversationStart timestamp in your local time zone. It is a bit more work but it is the only way to avoid the ‘Midnight Boundary’ issues!