Is there a hard rule on when to hit the live conversation endpoint versus the analytics one? We’re building a dashboard that needs to show agent status and current call metrics. My team has been using GET /api/v2/conversations to pull the list of active sessions, then iterating through the IDs to get participant details. It works fine for low volume, but the latency is creeping up during peak hours (around 4pm Sydney time).
I noticed the analytics endpoint GET /api/v2/analytics/conversations/realtime/query exists. The docs say it’s for ‘analytics’, which sounds like it’s for historical reports. Can I actually use it for real-time operational data? I tried a test query with this payload:
{
"interval": "5m",
"groupBy": ["mediaType"],
"aggregates": [
{
"name": "conversation/active",
"type": "count"
}
]
}
It returns data, but it feels like I’m abusing the endpoint. The live API gives me specific participant objects I can link to our UI, while analytics gives me buckets. If I need granular data (like specific agent names in a queue), do I have to stick with the slow live endpoint, or is there a trick to get that from analytics without hitting rate limits? We’re on Node.js, using the standard axios library for requests.