Root Cause Analysis
ArcIn Root Cause Analysis correlates metrics, traces, logs, deployments, and topology to identify the single most likely cause of an incident — and presents the evidence behind it, so you validate in seconds instead of investigating for an hour.
How it works#
When a service degrades, ArcIn walks the causal entity graph outward from the symptom, scoring each connected entity on timing correlation, dependency direction, and anomaly strength. Recent deployments, config changes, and infrastructure events are weighted heavily because they are the most common triggers.
What you see#
The RCA panel shows a ranked cause with a confidence score, the evidence chain (e.g. "deploy v412 → GC pause on order-api → checkout latency"), and a one-click path to the offending trace, log line, or change. No query writing required.
Design principle: ArcIn always shows why it reached a conclusion. You can expand every step of the evidence chain and disagree with it.
Acting on the result#
From the RCA you can trigger a runbook (see Auto Remediation), open an incident (Incident Management), or push the finding to Slack/ServiceNow. Feedback ("this was / wasn't the cause") tunes future analyses via IntelliTune.