cometchat-analytics
Installation
SKILL.md
Ground truth: shapes are FETCHED from the live docs —
{DOCS_BASE}/rest-api/metrics/usage-metrics,/rest-api/metrics/message-metrics,/fundamentals/webhooks(+/rest-api/management-apis/webhooks/overview),/notifications/logs,/rest-api/calls-apis/get-call, and on-prem/on-premise-deployment/docker/monitoring.DOCS_BASE = https://www.cometchat.com/docs; append.md. Verify each metric field / event name against the docs before charting it; don't invent a metric.
Use this skill when
You need visibility into chat: product analytics (DAU, messages, retention), an ops/observability signal (delivery, errors, latency), or feeding chat events into an existing warehouse/monitoring stack.
Three layers of visibility
| Layer | Source | Use it for |
|---|---|---|
| Aggregate metrics | Metrics REST API (usage + message) | dashboards, MAU/DAU, message volume, growth |
| Real-time events | Webhooks | stream every message/user/group/call event to your pipeline |
| Client health | your error monitor + the kit's error boundary/listeners | init/login failures, render errors, UX latency |
1. Aggregate metrics (REST, server-side)
- Usage metrics (
{DOCS_BASE}/rest-api/metrics/usage-metrics) — active users and usage over time. - Message metrics (
{DOCS_BASE}/rest-api/metrics/message-metrics) — message volume/counts. - Call metrics — per-call detail incl. participant metrics via
{DOCS_BASE}/rest-api/calls-apis/get-call. Call these from a server job with the REST API Key (never client-side), on a schedule, into your BI tool. Fetch the exact query params + response fields from each page before building the report.