Palanor publishes a live operational status page at palanor.com/status. It reads the platform's own health and reports it back in plain language. There is no manual toggle — the page reflects what is actually happening on the platform right now.
What the page shows
The headline at the top synthesizes every row into one of three states:
- **All systems operational** — every surface is green
- **Some systems degraded** — at least one row is amber (partial functionality)
- **Active incident** — at least one row is in outage state
Below the headline, one row per surface. Each row carries a state dot, a state label, and a one-line context note ("12 readings in last 36h," "Last run · today").
The surfaces
The page covers what stewards depend on:
- **Marketing site** — palanor.com itself
- **Application** — the steward-facing product at app.palanor.com
- **Custom Indices compute** — the nightly job that computes the twelve Palanor Indices
- **Schemas scoring** — the quarterly job that scores the ten Schemas
- **News ingestion** — the hourly pull from twelve newsrooms
- **News ranking** — the LLM pass that ranks coverage against your business profile
- **FRED economic-signal sync** — the FRED data ingest
- **Data plane** — the database and storage subsystem
- **Numen agent** — the LLM the agent reads through
- **Email delivery** — Numen brief delivery via Resend
- **Lattice Canvas persistence** — the persistence layer for graph-authored scenarios
What the states mean
**Operational (green)** — the surface is running on its expected cadence, with output that lands in the database / inbox / dashboard as it should.
**Degraded (amber)** — the surface is partly working. Could be: a vendor returning slower-than-expected responses, a cron that ran but produced fewer outputs than typical, an integration that requires a configuration step on your side. Steward-facing functionality often still works in amber; the page errs on the side of being honest about partial states rather than hiding them.
**Outage (orange / coral)** — a surface is not producing output. Investigation underway.
What the page does not tell you
By design, the page does not expose:
- Vendor names — when a vendor is degraded, the row reads "Numen agent — degraded," not the vendor's brand
- Hour estimates on in-flight work — that lives in the internal portal
- API costs, token usage, infrastructure costs
- Engineering bugs not visible to stewards
The full operational dashboard with all of that is the internal Portal — gated to the Palanor team.
Why this page exists
The point of an instrument is that you can trust it. Status is the proof.