Dashboard guide
Molar Guard has two customer-facing surfaces:
| Surface | URL | Purpose |
|---|---|---|
| Product landing | guard.molar.it | Marketing, positioning, feature overview |
| Operator dashboard | app.molar.it | Day-to-day PR gate, monitoring, Mender, settings |
Self-hosted Guard stacks redirect to the all-in-one dashboard when configured. Use app.molar.it for live /v1 data, SSE, and Mender controls.
Note:
Combined dashboard at app.molar.it is the recommended operator UI. Guard API calls go through an authenticated dashboard API proxy (/api/guard/* → Guard :4000/v1/*).
Navigation map
In the all-in-one dashboard, Guard surfaces live under the Data and Agents sections:
Overview (/)
├── Production guard (/guard) ← monitoring summary + PR gate strip
├── PR gating (/guard/pr-gate) ← recent PR checks
├── Runs (/runs) ← all run types
├── Scenarios (/scenarios) ← catalog (platform-wide)
├── Incidents (/guard/incidents) ← ack / suppress
├── Shadow-diff (/guard/shadow-diff) ← drift review (per run)
├── Auto-fix (/fixbots) ← Mender inbox
├── Settings (/settings) ← GitHub, API keys, synthetic safety
└── Audit log (/audit)
Guard-specific pages are gated by your org's guard entitlement. If you see an upsell screen, enable Guard on your plan or contact support.
Production guard (/guard)
At-a-glance production health:
- Live monitoring grid summary (scenarios × regions) from
GET /v1/monitoring/grid - Overall status banner from
GET /v1/status - Synthetic safety strip and PR gate compact panel
PR gating (/guard/pr-gate)
Recent PR gate activity:
- Latest PR runs and check conclusions
- Link through to run detail for logs and artifacts
Runs (/runs)
Universal explorer for every Guard run — PR, scheduled, manual, shadow_diff.
List view
Filter by type, status, repo, and date range. Columns include scenario, type, duration, region, and PR/commit metadata when present.
Detail view (/runs/{id})
| Section | Content |
|---|---|
| Live execution | SSE log tail via GET /v1/runs/:id/logs |
| Failure bundle | Stack, screenshot, video, HAR, console (signed artifact URLs) |
| Trigger metadata | Base/head SHA, affected set when PR run |
| Mender panel | Classification, confidence, preview, Apply fix (suggestive mode) |
Actions: Rerun (via API or MCP), open shadow-diff review when run_type=shadow_diff
Scenarios (/scenarios)
Platform scenario catalog shared with Cartographer. For Guard-specific schedule and shadow_prod settings, use the CLI, REST scheduled_checks API, or scenario frontmatter.
Monitoring (/guard/monitoring)
Matrix: rows = scenarios, columns = regions. Each cell shows current health status from the live Guard API. Drill down to recent runs and incident links.
Schedule CRUD is available via CLI (molar-guard schedule) and REST (/v1/scheduled_checks). A dedicated schedules editor in the dashboard.
Incidents (/guard/incidents)
- Filter: open / acknowledged / suppressed / resolved
- Types:
consecutive_failures,failure_rate,shadow_diff,global_outage - Actions: ack, suppress, link to root run artifacts
Shadow-prod review (/guard/shadow-diff)
Side-by-side prod vs Clone observation diff for a run. Approve, reject, or snooze expected vendor changes.
Auto-fix / Mender (/fixbots)
Inbox + attempt detail. See Mender auto-fix.
- Preview patch:
GET /v1/mender/:id/preview - Apply (suggestive → real PR):
POST /v1/mender/:id/apply - Reject with reason:
POST /v1/mender/:id/reject
Settings
| Area | What you configure |
|---|---|
| GitHub | App install, repo selection, required vs advisory |
| API keys | Create/revoke keys for CI and API |
| Synthetic safety | Analytics filter ack, middleware checklist |
| Mender | Mode, forbidden paths, token budget |
| Team / SSO | Members and roles (platform-wide) |
| Billing | Plan tier and usage |
Configure alert channels (Slack, PagerDuty, webhooks, etc.) in each scheduled check's alert_policy JSON today. A centralized notifications UI.
Status page
Public health JSON is available at GET /v1/status. Hosted guard.molar.it/status pages.
Keyboard and command palette
In combined dashboard mode, use the platform command palette (⌘K) to jump to runs, incidents, or scenarios.
Next
- Quick start — local dashboard at
:3000 - GitHub App setup — connect GitHub from Settings
- Production monitoring — monitoring grid deep dive
- Security — roles and who can change settings