> ## Documentation Index > Fetch the complete documentation index at: https://docs.zentien.io/llms.txt > Use this file to discover all available pages before exploring further. # Correlation > See when infrastructure outages and Salesforce error spikes happen at the same time. ## Overview Correlated incidents connect the dots between your infrastructure and your Salesforce errors. When an external service goes DOWN and a spike in Salesforce errors occurs around the same time, Zentien creates a correlated incident so you can see them together in one place. This answers the question: *"Is the Salesforce error spike happening because the payment gateway is down?"* ## How correlation works Zentien scans for overlapping windows: 1. An infrastructure check transitions to DOWN 2. Within a configurable time window, a spike in Salesforce errors for the same org is detected 3. A correlated incident is created linking the infrastructure outage and the error spike The correlation is displayed on the **Correlation** page and counted in the **Correlated Incidents** card on the Dashboard. ## Reading the Correlation page Each correlated incident shows: * **Infrastructure check** that went DOWN — name, type, and when it failed * **Downtime window** — start and end of the outage * **Salesforce errors** — list of error groups that spiked during the outage, with occurrence counts * **Timeline** — visual overlap of the infrastructure outage and error spike ## Interpreting results Not every correlated incident means the infrastructure outage *caused* the Salesforce errors — correlation is not causation. Use it as a starting point: * If errors are in integration classes (callout wrappers, API clients) and the failing check is the target API, the cause is likely the outage. * If errors are in unrelated classes, it may be coincidence. * Use the **Callout Monitor** on the Pulse page to see which endpoints were being called and whether they were failing. ## Dashboard widget The **Correlated Incidents** widget on the Dashboard shows the count of active correlated incidents. Click it to go to the full Correlation page. ## Next step Monitor org health — jobs, limits, callouts, and anomaly detection.