Production truth
Sentry is production-verified only after the read-only verification script proves env, source maps, monitors, alerts, and MCP reads against the live org.
This page reflects local production readiness for Aweb's Sentry integration. It does not claim live verification until the production verification marker and Sentry smoke evidence exist.
This shell inherits authenticated page-route posture and is not exposed as a public marketing surface.
Runner registration and connector mutation stay blocked; signed Mission approval issuing, sandbox eval summaries, cost visibility, and operator sandbox interrupts are live.
Published /api/v2/os9 fleets, runners, approvals, and mission contracts use route-level auth and explicit OpenAPI registration; broader OS9 APIs remain closed.
production verified
Configured, deployable, approval-gated, or deliberately disabled.
$5000 credit basis; logs remain gated by default.
No local readiness blockers.
Sentry is production-verified only after the read-only verification script proves env, source maps, monitors, alerts, and MCP reads against the live org.
Each row maps a Sentry product surface to the Aweb governance state.
Client, server, edge, and global error capture are code-wired.
Release, org, project, and auth token are present for upload on deploy.
Trace/profile capture is SDK-wired with env-controlled sample rates.
Replay is masked by default with low full-session sample and 100 percent error-session capture.
Sentry Logs are enabled for console error/warn only and pass through redaction.
Next/Sentry automatic Vercel Cron monitor instrumentation is enabled, but live monitor inventory must be verified.
Read/debug Sentry tools route through MCP Warehouse policy.
Seer and use_sentry remain high-risk tools and require explicit governed approval.
Aweb budget target or spend-alert confirmation is present.
These are the non-negotiable controls for using the credit without creating a noisy or unsafe observability surface.
SDK config keeps sendDefaultPii false.
Replay text and media stay masked unless a governed incident review allows otherwise.
Logs are code-wired for error/warn only and should stay off until spend alerts are verified.
Seer, Autofix, and use_sentry remain high-risk MCP actions behind approval.