@genesyscloud/px-engine is absolutely mangling the analytics payload again. I’ve got a custom desktop widget polling the queue observations query every 20 seconds to keep the agent dashboard fresh. Ran perfectly in dev, pushed to prod yesterday, and suddenly the network tab is just a sea of 400s. The request body is completely standard—time window locked to one hour, granularity set to interval with a 30m bucket. Nothing fancy.
Server logs are doing jack all. Just this exact payload coming back on every single poll:
{
"errors": [
{
"code": "invalid_request",
"message": "Requested granularity interval exceeds maximum allowed duration for the specified date range.",
"path": "groupBy.granularity"
}
]
}
@genesyscloud/px-engine’s internal fetch wrapper is silently auto-injecting ?includeSummary=true without logging it. That flag is completely breaking the interval validator on the backend. I can strip the query param manually and the call works, but that’s a filthy workaround I’m not shipping. We’re on @genesyscloud/px-engine version 2.14.3. @genesyscloud/px-engine middleware is definitely intercepting and mangling the analytics payload somewhere in the chain. Anyone else hitting this with the embeddable SDK or screen pop configs?