Follow the lifecycle of a
critical outage

A frustrated customer first encounters the problem and decides to report it to support.

The support request is read and the agent creates a Jira ticket for the issue. Meanwhile, more users are affected and social media complaints start appearing.

The ticket waits for triage while the support team handles other requests. No one knows the severity yet.

Support escalates to engineering. An incident is created, but the damage to customer trust is already done.

Someone remembers to update the status page. Customers have been in the dark for hours.

Developer manually searches through Datadog, CloudWatch, and application logs trying to find the root cause.

Without clear traces, the team tries multiple potential fixes. Each deployment is a roll of the dice.

After hours of downtime, the fix works. Status page manually updated. Post-mortem conclusion: "We need better monitoring."

Uptime monitors on the network layer detect it is no longer returning expected status codes from multiple global locations.

Intelligent alerting routes the notification to the right team channel with full context: screenshots, traces, and error details.

API checks on core workflows that validate response payloads and detect schema violations, authentication failures, or degraded performance start to fail.

Browser checks running real Playwright tests detect that real user flows like login and checkout are failing.

An incident is declared and created in Rootly with full context of the outage to kick off the right incident response process.

Public status page automatically reflects the incident, keeping customers informed without manual intervention.

Rocky AI correlates traces, logs, and check results to identify the root cause and suggest fixes with code examples.

Fix deployed, checks pass, status page auto-updates to "Operational". Full incident timeline captured for post-mortem.