← Modfolio Connect

Modfolio Connect — Status

Independent live probes against the login.modfolio.io / account.modfolio.io production endpoints. Page rendered at ; per-row state refreshes client-side every 60 seconds.

Overall

Checking probe results…

Services

ServiceTargetCurrentLatency
JWKS — public signing keys
https://login.modfolio.io/.well-known/jwks.json
99.95%checking…—
OIDC Discovery
https://login.modfolio.io/.well-known/openid-configuration
99.95%checking…—
Auth issuer health
https://login.modfolio.io/health
99.9%checking…—
SCIM Service Provider Config
https://login.modfolio.io/scim/v2/ServiceProviderConfig
99.5%checking…—
Account dashboard
https://account.modfolio.io/health
99.5%checking…—

Service-level commitments

These are the customer-promised Tier-1 SLOs measured on a 30-day rolling window. The live "Services" table above reports endpoint reachability — the table below is the behavioural contract being measured against. Source of truth: docs/slo.md in the open-source repository.

ObjectiveTargetMeasurement
Token issuance success rate99.9%POST /sso/token 2xx / total
Token issuance p95 latency< 500 msPOST /sso/token p95
JWKS endpoint availability99.95%GET /.well-known/jwks.json 2xx / total
JWKS endpoint p95 latency< 100 msGET /.well-known/jwks.json p95 (cache-hit dominant)
Login flow completion rate99.5%authorize → token round-trip success / authorize total
Login flow p95 latency< 3 send-to-end browser → final session cookie set
Session verification success rate99.9%GET /api/auth/get-session 2xx / total

Error budget for a 99.9% target on a 30-day window is ~43 minutes; for 99.95% it's ~22 minutes; for 99.5% it's ~3.6 hours of degraded behaviour. Burn-rate alerting follows the SRE multi-window approach (1h/6h/1d/3d windows at 14.4×/6×/3×/1× burn rate).

Architecture

Modfolio Connect is edge-first by policy: every component below runs on Cloudflare's global network. No regional VMs, no managed control plane outside CF. When a probe above degrades, the responsible primitive is named here.

Recent incidents

No public incident posts yet. Past incidents are recorded internally at docs/incidents/ in the source repository and summarized publicly here once the postmortem is published. The hand-off to an independent statuspage vendor is tracked in DR plan §S4; until that lands, treat the absence of an entry here as "no events visible to the platform team," not "no events at all."