Scans
A scan is an on-demand sweep of a single connection: a cloud account or an observability integration. An agent walks the resources behind the connection, reads their metrics, logs, traces, and configuration, and writes a report with a summary, key risks, advisories, and follow-up questions.
Start a scan
Polylane offers a scan right after you connect a cloud account or an observability integration. Click Run scan on the confirmation page that follows the connect flow. Other integration types skip the scan and get starter questions instead.
The scan runs on Polylane's servers and takes about a minute. It continues if you leave the page, and the report lives at its own URL. While it runs, the scan page streams live progress. A cloud scan moves through mapping your resources, reading metrics, logs, and traces, then analysing and writing up. An integration scan reads your integration data, then analyses and writes up. Below the checklist, a feed shows the steps and readings pulled from your account as they arrive.
A scan ends in one of two states: ready or failed. A failed scan shows a Run the scan again button that takes you back to the connect confirmation page.
What the report contains
- Digest and summary: a one-to-two sentence digest, then a full report body.
- Key risks: the few findings worth acting on, usually about three and sometimes none when the account looks healthy. Each risk has a title, detail, a severity of low, medium, or high, and the affected resources. Every risk has an investigate button that opens a thread with a prompt like:
Investigate this risk on my AWS account: <title>. <detail> The affected resources are: <resources>. Check whether it has already shown up in metrics, logs, traces, or issues, and recommend what to do about it. - Advisories: deterministic best-practice findings computed from the context graph at scan time, grouped into observability, resilience, security, data protection, and lifecycle. See Advisories.
- Signals: extra tabs appear when the scan captured them. Metrics show current, minimum, average, and maximum values with a time series per metric. Logs show volume by severity plus the top and newly appeared log templates. Traces show trace digests.
- Follow-up questions: starter questions generated from the scan. Clicking one opens a thread scoped to the scanned account or integration, and you can type your own question inline.
- Topology: cloud scans embed the topology of the scanned account, updating live while the account keeps syncing.
The Copy button exports the digest, summary, metrics, advisories, and risks as Markdown.
Where scans appear
Every workspace has a scans page listing all reports, sortable by account, provider, risks, metrics, status, and creation time. Each report has its own page.
The sidebar shows a transient entry while a scan is active: Scan running during the sweep, then Scan report ready until you open the report once. The ready entry shows for up to 24 hours after the report was generated.
Share a scan report
Reports are visible to your workspace by default. The share modal supports two visibility levels: workspace and public. Setting a report to public creates a share link of the form:
https://console.polylane.com/{workspace-slug}/scans/p/{scan-id}
Anyone with the link can view a public report without signing in. The public version strips the account and region identifiers, the connection identifier, and the creator, and omits the follow-up questions and the embedded topology. Set the report back to workspace visibility to revoke the link.
Try it from your coding agent:
Advisories
Per-resource recommendations. Misconfigurations, resilience risks, and observability gaps that agents find on the resources in your topology.
Overview
When something needs working, an agent runs the investigation with your real tools and shows the receipts. Threads, investigations, and escalations are the surfaces.