APIN

warming up the server…

first visit may take up to 30s · subsequent visits are cached

APIN logo
APIN field laboratory
← Disease detector

System Status

——:—— IST

Live health of the APIN plant-disease service. This page refreshes itself every 30 seconds.

The status feed could not be reached just now — the figures below may be a little stale. Retrying automatically.

Checking system status…

Contacting the status feed…

90-day uptime
Diagnoses served
Median check
Compute device

External Availability

ground-truth uptime measured from outside the container
24h uptime (URL-reachable)
7-day uptime (URL-reachable)
90-day uptime (URL-reachable)
p95 latency (24h)
probes / 24h
incidents (24h)

URL-reachable — HTTP 2xx + valid response shape + latency under SLO. This is what users see. Fully-healthy (smaller number underneath) additionally requires every component to report up AND no resource alerts — the stricter engineering SLO. The two numbers will diverge during ops-only incidents (high memory, degraded subsystem) that don't affect responses.

Waiting for external probe data — the GitHub Action + Cloudflare Worker write to Turso every 1–2 minutes; the chart below fills in as days accumulate.

Probe sources (last 24h)

independent vantage points; disagreement reveals network-path issues
Waiting for first probe results…

Recent external incidents

failed probes from the last 7 days
No external incidents recorded yet.

Component Health

90-day history · one cell per day · inside-the-container view
Loading component health…

Recent pulse

recent heartbeats · hatched marks are gaps
Waiting for heartbeat history…

Past incidents

component faults in the 90-day record
Loading incident history…

Past interruptions

monitor gaps in the last 48 hours
Loading interruption history…
Operational Degraded Down No data · service asleep

Uptime bars fill in over time — a brand-new deployment shows mostly “no data” until 90 days of heartbeats accumulate. Past incidents are component faults recorded while APIN was running. Past interruptions are gaps in the heartbeat stream — a crash, a redeploy, or the free Space going to sleep, when nothing was alive to record a heartbeat. A total outage of the Space is also tracked by a separate external monitor.