Vendor status

The Vendor Status panel shows a live status pill for every API you depend on, updated every 60 seconds — independent of your SDK traffic.

The panel appears on the Overview dashboard below the error rate chart. It runs whether or not requests are currently flowing, so you get signal even on quiet days.

Status verdicts

Each vendor gets one of four verdicts, resolved in priority order:

VerdictMeaning
OutageA critical or major incident is active on the vendor's status page.
DegradedA minor incident is active on the vendor's status page.
UnreachableThe latest probe failed — TCP/TLS connection refused or timed out.
OperationalNo active incidents and the last probe succeeded.

Degraded components from a vendor's status page are listed below the pill as informational context. They do not change the verdict color on their own — only active incidents do.

How probes work

Probes run every 60 seconds via TCP/TLS connection checks across all built-in vendors: Stripe, OpenAI, Anthropic, Twilio, Resend, and GitHub. They operate independently of SDK traffic — status is always current regardless of how many requests your app is making.

Custom vendors added via extra_vendors are not probed. The panel only shows built-in vendors that have appeared in your traffic in the last 30 days.

TLS certificate warnings

When a vendor's TLS certificate has fewer than 14 days remaining, a warning is shown inline below the status pill. This is separate from the status verdict — a vendor can be Operational while still showing a certificate warning.

Active incident summaries

Detected incidents are listed below each vendor's pill with their title and severity level. The list is sourced from the vendor's public status feed and refreshed on every probe cycle.