Skip to content
statance
MenuClose

An alert starts
the investigation.

Bring checks, server metrics and processes together. Let AI investigate likely causes, then inspect the evidence behind the explanation.

Start monitoring ↗

checkout.api

HTTP 502 · 14:32

Incident investigation

Disk filled before
the endpoint failed.

On web-01, disk usage reached 100% at 14:31. The first HTTP 502 was recorded a minute later, with failures from three regions.

A log-compression process was active around the same time. Check available disk space and the files being written before restarting the service.

Suggested next step

Free disk space, confirm the application can write again, then add a disk alert above 85%.

Ask a question.
Follow the evidence.

Use the assistant to explore your infrastructure in plain language. Or connect an MCP-compatible tool to your Statance data.

Connect through MCP ↗
Choose a question to explore the connected monitoring data.

Worth knowing.

Which plans include automatic investigation?

Pro and Business include AI incident investigation. Free includes a limited AI assistant.

Can I use my own assistant?

Yes. Statance provides a remote MCP server and REST API, authenticated with your account’s API key.

Does the explanation prove the root cause?

It gives you a likely cause and evidence to investigate. Review the underlying checks and metrics before making operational changes.

Uptime monitoring.
With super powers.

Start monitoring ↗

Follow the evidence, not just a plausible explanation.

Statance’s interactive assistant can query server metrics, processes, monitors, checks, incidents, alert rules and security findings for your account. It is instructed to fetch real monitoring data and avoid inventing numbers. It can also create, pause or resume monitors; its instructions require an explicit user request for those changes. It does not have a tool that opens an SSH session or executes arbitrary commands on your server.

Automatic first-look notes have a defined evidence scope.

For a server incident, the current automatic investigation gathers server metrics and process data. For a monitor incident, it gathers the monitor’s recent checks. It also gathers relevant security findings, their scan health and recent related incidents, then asks the model for a short note using that evidence, with tool calls disabled during generation. A monitor incident is not automatically mapped to a particular server and enriched with that server’s resource history.

Use the assistant to ask follow-up questions across the data you have connected. A suspected disk issue is a hypothesis to verify against mount usage, logs and changes, not proof produced simply by a coinciding spike.

Security context makes the conversation more useful.

On eligible plans, security findings can help you inspect vulnerable or outdated software, end-of-life versions, suspicious processes and website scan findings alongside operational metrics. This supports prioritization and investigation. It is not a guarantee of compromise detection, antivirus replacement or automatic remediation.

Ask: “What do the failed checks actually show?”, “Which server has high disk usage?”, and “Which security findings need attention on this host?” Review the evidence and permissions before acting. Missing telemetry, stale inventory or incomplete checks should remain visible as uncertainty.

Read the assistant documentation · Explore security monitoring · See a reproducible probe investigation