Role: SA · Window: last 24h · Generated: 2026-08-23T11:37:48.472Z · Auto-refresh: 60s
Workers tracked
80
Total requests (24h)
35.6k
Overall error rate
0.00%
AI Gateway cost (24h)
$0.0000
How to read this: Summary stats give you the shape of the day. The error rate is the first thing to scan — anything over 5% on a Worker handling real traffic is worth a click-through. AI Gateway cost trending up faster than request count usually means a model swap or a cache-miss pattern; investigate.
Workers — by request volume (24h)
Worker
Requests
Errors
Err %
CPU p50
CPU p95
mycelia-api
6.8k
0
0.00%
2543.00ms
5139.00ms
heartbeat-watcher
3.5k
0
0.00%
465.00ms
1150.00ms
cistern
2.8k
0
0.00%
2072.00ms
4761.00ms
princess-polka-dottie
861
0
0.00%
400.00ms
767.00ms
velovigil-fleet
737
0
0.00%
345.00ms
1253.00ms
daemon
730
0
0.00%
433.00ms
702.00ms
northwoods-daemon-network
712
0
0.00%
351.00ms
10788.00ms
google-edge
641
0
0.00%
2015.00ms
3434.00ms
mistral-worker
559
0
0.00%
1087.00ms
1793.00ms
gemini-worker
523
0
0.00%
1163.00ms
1700.00ms
fleet-relay
521
0
0.00%
372.00ms
892.00ms
fleet-dashboard
517
0
0.00%
2963.00ms
4462.00ms
mycelia-watch-cf
515
0
0.00%
1214.00ms
1901.00ms
perplexity-worker
512
0
0.00%
1021.00ms
1611.00ms
mirror-worker
487
0
0.00%
1074.00ms
1732.00ms
lightboard
479
0
0.00%
431.00ms
1270.00ms
northwoods
468
0
0.00%
284.00ms
487.00ms
brook
467
0
0.00%
631.00ms
976.00ms
smallest-loop
355
0
0.00%
550.00ms
3696.00ms
cf-heartbeat-watcher
354
0
0.00%
496.00ms
2070.00ms
overnight-builder
349
0
0.00%
412.00ms
979.00ms
appsec-gate
324
0
0.00%
254.00ms
1001.00ms
fleet-error-watch
320
0
0.00%
1456.00ms
1894.00ms
bivouac
281
0
0.00%
541.00ms
1515.00ms
edison-mirror
273
0
0.00%
416.00ms
788.00ms
librarian
271
0
0.00%
564.00ms
5269.00ms
pai-attest
255
0
0.00%
401.00ms
609.00ms
finops-do
254
0
0.00%
595.00ms
836.00ms
ingest-cisa-kev
253
0
0.00%
329.00ms
917.00ms
sa-dashboard-kit
252
0
0.00%
278.00ms
376.00ms
Pattern to watch: p95 climbing while p50 stays flat is usually downstream contention (D1, KV, external API). p50 and p95 both climbing means the Worker itself is slowing down. CPU time is wall-clock for the Worker only; downstream wait time isn't here, which is why latency anomalies can hide in plain sight.