Dashboard
The Dashboard is the screen you land on after signing in (the sidebar's Dashboard item, URL /). It's a live "operations wall" that shows what's happening right now across your tenant: incidents coming in, how triage is performing, and what the AI agents are doing.
What's on the dashboard
The dashboard brings together a few kinds of information:
- Key performance indicators (KPIs): at-a-glance numbers such as:
- Open incidents: how many incidents are currently open.
- P1 / critical: count of the highest-priority incidents.
- SLA 15m: performance against your service-level target.
- Triage p50: the median time triage takes.
- LLM tokens: AI usage.
- HITL backlog: how many actions are waiting for human approval (see Human-in-the-Loop).
- Impact metrics: headline figures that show the value CloudPeek is delivering, such as human hours returned to the team, reduction in time-to-triage, and triage success rate.
- Live activity: a real-time view of incidents and agent activity as they happen.
The exact tiles you see depend on your data and configuration. On a new instance most numbers start at zero and fill in as alerts arrive and triage runs.
The "no AI provider" banner
If no AI model provider is configured yet, the dashboard shows a banner prompting you to set one up. Triage and investigations can't run without it. Fix this in Settings → Model Providers (see the Quickstart).
Getting started checklist (onboarding)
New instances include an onboarding dashboard (/onboarding) with a setup checklist that walks you through standing CloudPeek up:
- Connect first source: add a tool that provides alerts.
- Ingest first alerts: pull alerts in.
- Run first triage: let triage analyse an incident.
- First investigation: run an AI investigation.
- SLA target set: define your service-level target.
- Review HITL backlog: clear any actions awaiting approval.
- Add 2nd source: connect another tool.
Each item links straight to the screen where you complete it.
Using the dashboard day to day
- Scan the KPIs to see whether anything needs attention (a rising P1 count, a growing HITL backlog).
- Jump from a live incident straight into its detail view to work it (see Incident Management).
- Use the impact metrics when reporting to stakeholders on how the team is performing.
Related screens
- Analytics: deeper, historical dashboards for performance, agents, cost (FINOPS) and audit.
- Incident Management: the full incident queue behind the KPIs.