Introducing AI Root Cause Analysis
MonSPHERE now correlates signals across your entire stack to surface the real cause of incidents in seconds, not hours.
When an alert fires, the hard question is never "what broke" — it's "why." A CPU spike on one host, a latency increase on a downstream service, and a burst of 5xx responses might be three symptoms of one root cause, or three unrelated problems. Today, most teams answer that question by opening five browser tabs.
What's real today
MonSPHERE's incident pipeline already correlates related signals automatically, because every alert, incident, and audit event flows through the same internal event bus. When alert-engine fires multiple related alert rules within a short window, incident-service groups them onto one incident timeline instead of opening five unrelated tickets — so the correlation you'd otherwise do by hand across dashboards happens structurally, at the data layer.
curl https://api.monsphere.com/api/v1/incidents/:incidentId \
-H "X-Api-Key: msk_live_..."
# { "id": "...", "relatedAlerts": [...], "timeline": [...] }What's still ahead
Full AI-assisted root cause analysis — automatically ranking which of several correlated signals is the likely cause, not just grouping them — is on our roadmap and not shipped yet. We'd rather tell you honestly where the line is today than let a headline overstate it. See the Roadmap for status.
Why this approach
Grouped, correlated incidents are useful on their own even before any ML ranking is layered on top — an on-call engineer looking at one incident with five related alerts and a shared timeline is already faster than five separate pages with no relationship between them.