Hey everyone,
Something’s really messed up with how contest entries are being calculated in our new “Top Talker” leaderboard. We launched it last week to boost handle time - figured a little friendly competition would motivate the team, and it was working, until yesterday. Now the entry counts are all over the place. It’s impacting morale, and honestly, HR is asking questions.
The contest itself is pretty simple - 1 point for every minute of talk time, using the AudioTalkTime metric. We’ve got it tied to the agent’s profile, and entries are incrementing the totalPoints custom field. The flow looks straightforward - a data action that queries the user’s metrics, filters by metricId equal to AudioTalkTime, then calculates the total minutes and updates the profile. It’s been working for a week, so I’m stumped.
We’re seeing some agents with drastically inflated scores - like one agent with 1200 minutes in the last 24 hours when their average is usually around 300. Other agents have zero points, despite having calls. The weirdest thing is, when I look at the individual metric data in the Agent Desktop, it shows the correct talk time, but it’s not being reflected in their contest entry total.
Is anyone else running into this? Is there some sort of caching issue with the metric API? We’ve tried restarting the flow, flushing the cache (though I’m not sure that actually does anything), and even temporarily disabling the contest to see if that resets things, but no luck. We’re on Genesys Cloud v24.4.1. Any thoughts? Honestly, if I have to explain to Brenda in HR again why the leaderboard is broken, I might lose it. I’ve got a temporary workaround - manually updating the profile fields via the API, but that’s not a solution.
Here’s what we’ve checked:
- Tenant version: v24.4.1
- Flow version: 2.1 (last updated 7/18)
- Data action endpoint: User metrics query
- Metric ID:
AudioTalkTime - Time range: Last 24 hours
- SDK version: Using the standard Genesys Cloud SDK within the Architect flow.
- We’ve reviewed the documentation - it’s not exactly clear if there are rate limits on the metric endpoint.