All docs
Getting started
Your first five minutes
This page tells you what to expect in the minutes after the agent connects, where to look on the server page, how to act on the first findings, and the two short setup steps that make sure a critical alert reaches a human.
Within one minute: the server card
As soon as the first heartbeat lands, the server appears on the fleet grid. The card shows an online dot, the name you gave it, and columns for Health (0–100), Security (0–100), CPU, Memory and Services. Click the card to open the server page.
Within a few minutes: the score and findings
The health score and the first findings come from the first snapshot and are refreshed every 60 seconds. A few rules need more than one heartbeat before they can judge, for example rates that compare two samples, so the score can move a little during the first minutes. That is expected.
The Overview tab opens with Needs attention: critical and warning findings sorted critical-first, each with a plain-English cause and a fix. Info-level items sit below under Low-priority notes so nothing is hidden. For the full breakdown of how each point was lost, see Health score.
The server page tabs
| Tab | What you find there |
|---|---|
| Overview | Score, what needs attention, and the quick facts about the host |
| Activity | Logs clustered into patterns, and a Timeline that correlates what changed before an incident |
| Security | The security score with its breakdown, plus a Compliance sub-tab with SOC 2, PCI-DSS and HIPAA evidence |
| Automation | Remote actions, the terminal and command history |
| Deploy | Deploy and rollback for apps the agent found |
| Workloads | Services, Containers, Processes and Apps |
| Data stores | MySQL, Redis and PostgreSQL, shown only when they are present |
| Network | Ports and Domains |
| System | Disk, Files, Users, Packages and Scheduled jobs |
| Trends | History charts of the collected metrics |
Tabs for workloads, data stores and so on only appear when the agent actually found something to show.
Act on a finding
Each finding carries the command that fixes it. You can copy it and run it yourself over SSH. If you turn on remote actions for the server, the finding gets a Run button that executes the fix from the dashboard with a confirmation and a full audit record, and Fix with AI walks a harder problem to done: it proposes, explains, runs on your approval and verifies. Remote actions are off by default and are enabled per server. See Remote actions.
Ask the AI chat
Open Chat and pick a server. The empty state offers starter questions you can click:
- •Why is this server slow?
- •Which servers have SSL expiring soon?
- •Show me all critical issues across my infrastructure
- •What changed in the last 24 hours?
Answers are built from your live data and list the data sources used. See AI chat.
Do these two things today
1. Add a notification channel
Alerts are only delivered to channels you add. Until the workspace has one, critical alerts exist only inside the dashboard, and a red banner at the top says so: "N critical alerts are open and this workspace has no notification channel — nobody has been emailed or messaged." The banner links to Add a channel (1 min) and can be hidden for 24 hours. While the condition lasts, workspace owners also get a reminder email at most once a week. Add Slack, Telegram, email, a webhook, PagerDuty or OpsGenie under Account. See Alerts and channels.
2. Add an uptime monitor
The agent cannot report that the server is gone. An external HTTP check on your public URL, plus a Domain check for certificate expiry, catches a full outage from the outside. See Uptime monitors.
Add the rest of your servers
Repeat Install the agent for each host. The Free plan covers two servers; see Plans and billing for more.
Last updated 2026-10-01. Found a mistake? Email hello@getopservo.com.