Sentinel

Sentinel watches the workloads in your boundary, explains in plain language what is wrong with them, and tells you what to do about it.

• What Sentinel sends to the model — Which tenant data Sentinel sends to the LLM gateway, when log lines are included, and what is removed first.
• Findings — What a Sentinel finding contains, what its severities, categories and statuses mean, and when it closes, reopens and is deleted.
• How Sentinel works — The rule checks and analysis runs that produce findings, what Sentinel reads, what it sends to a language model, and how boundaries are kept apart.
• Sentinel — Sentinel watches the workloads in your boundary, explains in plain language what is wrong with them, and tells you what to do about it.
• Permissions — The actions that govern Sentinel, the roles that hold them, and why they must be assigned on the boundary.
• Sentinel on a resource — The Sentinel card on a resource's page, what turning monitoring on and off does, and the minimum severity setting for apps.
• The Sentinel page — What the Overview, Findings, Runs and Platform tabs of the Sentinel page show.
• Sentinel settings — The language model Sentinel needs, the settings that control its runs, rule checks, retention and log evidence, and the defaults.
• Work with findings — Change a finding's status, have Sentinel analyse it with the resource's logs, and report a suspected Stackship bug.