Ingestion
{{ verdict.word }}

{{ status?.day ?? '…' }} · {{ taskCount }} tasks · {{ groupCount }} feeds · refresh 2s · {{ status.branch }} · status stale {{ staleFor }} · dev → {{ apiBase.replace('https://', '') }}
api error: {{ fetchError }}

› HOST · SYSTEM GAUGES 01 / 03

The box the controller runs on — CPU, memory, the local data disk, and this daemon's own footprint. Sparklines cover the selected window (line = avg, band = per-bucket max, so bursts survive); the ring is in-memory and restarts empty with the daemon.

cpu load

{{ pct(sys.cpu?.percent) }}%

{{ sys.cpu?.vcpus ?? '–' }} vCPU · load {{ sys.cpu?.load_avg_1m ?? '–' }} / {{ sys.cpu?.load_avg_5m ?? '–' }} / {{ sys.cpu?.load_avg_15m ?? '–' }}

memory

{{ pct(sys.memory?.percent) }}%

{{ bytes(sys.memory?.used) }} / {{ bytes(sys.memory?.total) }} used

disk · {{ sys.disk?.path ?? '–' }}

{{ pct(sys.disk?.percent) }}%

{{ bytes(sys.disk?.used) }} / {{ bytes(sys.disk?.total) }} used

network · ↓ download

↑ {{ rate(sys.network?.sent_bps) }} · Σ ↓ {{ bytes(sys.network?.recv_total) }} since boot

daemon · ingests running

{{ sys.ingests?.running ?? 0 }} / {{ sys.ingests?.tasks ?? '–' }} tasks

rss {{ bytes(sys.process?.rss) }} · {{ sys.process?.num_threads ?? '–' }} thr · {{ sys.process?.num_fds ?? '–' }} fd · up {{ dur((sys.process?.uptime_s ?? 0) * 1000) }}

› TIMELINE · WHERE THE TIME GOES 02 / 03

Every execution of the day on a real time axis — one lane per task, its detect probes, ingests and stages. Each task keeps its color; the glyph + label tell which stage. A lane's swatch pings whenever its task does anything and breathes while a stage runs, so you can see who is working without hunting the right edge. Live view holds the last two hours and follows them; any zoom or pan takes control back.

span {{ spanLabel }} · following the live edge

lane

no executions recorded for {{ dateStr }} yet

loading…

› TASKS · WHAT LANDED 03 / 03

One tile per task, color-coded by feed. The hero is the sessions ingested on the selected day — every feed is T+1, so "waiting for today's data" is the normal resting state and says nothing; what a task actually drained does. Below it: the detect freshness, the backlog it still owes, and its pipeline stages.

ingested · {{ isToday ? 'today' : dateStr }}

detect

checked {{ ago(t.detect?.checked_at) }}

ingest

{{ t.ingest?.state ?? 'idle' }} · day {{ t.ingest.cursor.index }}/{{ t.ingest.cursor.total }} · {{ t.ingest.cursor.current }}