Xray
hourly rollups 1h 6h 24h 48h 3d 5d 7d 14d 30d ‹
Last 3 days▾
Ends now. Anything up to this cluster's 30-day retention.
Start and end times
Cancel
Interpreted in UTC (change it at the right of this bar). Earliest available: 2026-10-20 03:13:42.
› →|
updated 1d ago
Sample data. This database was generated by scripts/seed_dev_db.py — every service, deploy, person and figure on this page is synthetic. Names are sample.*@seed.invalid and image tags are prefixed SEED-. Nothing here happened.

Services 20

Per-service rollup over the last 3d. Click a service for its endpoint-level X-ray.

Neo4j fan-out per request
Calls (log) × queries/request — up-and-right is an N+1 suspect (amber ≥3, red ≥10 q/req)
0510151101001k10kcalls (log scale) →q/req ↑onboardingsvc · calls 1,862 · q/req 0.9studiosvc · calls 830 · q/req 2.3execsvc · calls 653 · q/req 12.6execsvcvirtual-assistant · calls 216 · q/req 3.7virtual-assistantdocsvc · calls 174 · q/req 3.0enrollsvc · calls 169 · q/req 2.8workplansvc · calls 149 · q/req 11.2workplansvcdashql · calls 82 · q/req 1.2tusker · calls 59 · q/req 2.1schedsvc · calls 32 · q/req 3.9schedsvcdocproc · calls 23 · q/req 2.4chat-agent · calls 18 · q/req 1.1ontologysvc · calls 13 · q/req 1.5patientsvc · calls 11 · q/req 2.9regsvc · calls 1 · q/req 1.0
Columns
Views
Export CSV
Service API calls ▾ Streams Err % Avg latency p95 latency Neo4j queries Background Unattributed Neo4j / req Write % Neo4j latency
onboardingsvc 1,862— 0.8% 46 ms 151 ms 2,858 1,178 41.2% 0 0.9 7.5% 5 ms
studiosvc 830— 0.0% 337 ms 963 ms 1,873 0 0.0% 0 2.3 4.1% 28 ms
execsvc 653— 4.9% 319 ms 1.55 s 8,420 221 2.6% 0 12.6 47.7% 10 ms
virtual-assistant 216⇄ 29 conns 0.0% 477 ms 2.90 s 3,511 2,598 74.0% 0 3.7 11.3% 13 ms
docsvc 174— 8.0% 659 ms 3.94 s 2,012 1,498 74.5% 0 3.0 6.2% 29 ms
enrollsvc 169— 0.6% 171 ms 446 ms 551 78 14.2% 0 2.8 13.4% 13 ms
workplansvc 149⇄ 10 conns 0.0% 289 ms 1.14 s 48,457 46,675 96.3% 0 11.2 9.9% 1 ms
dashql 82— 0.0% 1.13 s 3.99 s 100 1 1.0% 0 1.2 1.0% 11 ms
tusker 59— 5.1% 252 ms 2.05 s 2,994 2,869 95.8% 0 2.1 0.8% 10 ms
schedsvc 32— 9.4% 159 ms 440 ms 395 270 68.4% 0 3.9 3.0% 5 ms
docproc 23— 0.0% 163 ms 785 ms 12,818 12,763 99.6% 0 2.4 0.5% 19 ms
chat-agent 18⇄ 3 conns 0.0% 485 ms 5.50 s 24 0 0.0% 0 1.1 0.0% 81 ms
ontologysvc 13— 53.8% 104 ms 233 ms 24 5 20.8% 0 1.5 0.0% 3 ms
patientsvc 11— 36.4% 277 ms 1.10 s 32 0 0.0% 0 2.9 9.4% 53 ms
regsvc 1— 0.0% 46 ms 46 ms 1,295 1,294 99.9% 0 1.0 0.0% 4 ms
dashrsvc 0— — — — 103 103 100.0% 0 — 96.1% 45 ms
data_monitor 0— — — — 2,332 2,332 100.0% 0 — 39.9% <1 ms
finsvc 0— — — — 1,267 1,267 100.0% 0 — 0.2% <1 ms
notifsvc 0⇄ 226 conns 1.3% — — 343 165 48.1% 0 0.8 8.7% 4 ms
schema-migrate 0— — — — 8 8 100.0% 0 — 50.0% 2 ms

Neo4j queries = total in-window. Background = those with no matching HTTP request (pubsub / cron / worker paths) — e.g. a service with 8 API calls but thousands of queries is doing background work. Neo4j / req = foreground queries (total − background) ÷ API calls — the true per-request fan-out (high = possible N+1), no longer skewed by background load.

Who owns what — 31 workloads on nexus, from GitHub CODEOWNERS

ServiceOwnerBackup CODEOWNERS line
advpsvc Suresh BoddapatiSambasiva Bandarupalli tildabio/main:.github/CODEOWNERS /advpsvc/
chat-agent Sambasiva BandarupalliAhmed Aburaed tildabio/main:.github/CODEOWNERS /data-science/
dashql Vishal KaredlaTaylor Futral tildabio/main:.github/CODEOWNERS /dashQL/
dashrsvc Vishal KaredlaTaylor Futral tildabio/main:.github/CODEOWNERS /dashRsvc/
data-monitor Sambasiva BandarupalliAhmed Aburaed tildabio/main:.github/CODEOWNERS /data-science/
docproc Sambasiva BandarupalliAhmed Aburaed tildabio/main:.github/CODEOWNERS /data-science/
docsvc Prasad LimayeHarish Kumar tildabio/main:.github/CODEOWNERS /docsvc/
enrollsvc Suresh BoddapatiSambasiva Bandarupalli tildabio/main:.github/CODEOWNERS /enrollsvc/
execsvc Madhava MukkamalaShashank Agrawal tildabio/main:.github/CODEOWNERS /execsvc/
exportsvc Sambasiva BandarupalliAhmed Aburaed tildabio/main:.github/CODEOWNERS /data-science/
exportsvc-worker Sambasiva BandarupalliAhmed Aburaed tildabio/main:.github/CODEOWNERS /data-science/
finsvc Murali Krishna VelamatiSambasiva Bandarupalli tildabio/main:.github/CODEOWNERS /finsvc/
litellm nobody named — the leads (Ojas Kale, Sambasiva Bandarupalli) tildabio/ash-nazg:CODEOWNERS *
notifsvc Murali Krishna VelamatiSambasiva Bandarupalli tildabio/main:.github/CODEOWNERS /notifsvc/
onboardingsvc Prasad LimayeSambasiva Bandarupalli tildabio/main:.github/CODEOWNERS /auth*/
ontologysvc Sambasiva BandarupalliAhmed Aburaed tildabio/main:.github/CODEOWNERS /data-science/
patientsvc Suresh BoddapatiSambasiva Bandarupalli tildabio/main:.github/CODEOWNERS /patientsvc/
phase-2-nexus unowned — in no repository X-ray reads; the leads —
rag-agent Sambasiva BandarupalliAhmed Aburaed tildabio/main:.github/CODEOWNERS /data-science/
rag-agent-worker Sambasiva BandarupalliAhmed Aburaed tildabio/main:.github/CODEOWNERS /data-science/
regsvc Satish AshokSambasiva Bandarupalli tildabio/main:.github/CODEOWNERS /regsvc/
schedsvc Murali Krishna VelamatiSambasiva Bandarupalli tildabio/main:.github/CODEOWNERS /schedsvc/
schema-migrate Suresh BoddapatiOjas Kale tildabio/main:.github/CODEOWNERS /migrations/
studioaisvc Sambasiva BandarupalliAhmed Aburaed tildabio/main:.github/CODEOWNERS /data-science/
studiosvc Suresh BoddapatiMadhava Mukkamala tildabio/main:.github/CODEOWNERS /studiosvc/
studiosvc-infra-nexus unowned — in no repository X-ray reads; the leads —
swaggersvc nobody named — the leads (Ojas Kale, Sambasiva Bandarupalli) tildabio/main:.github/CODEOWNERS *
tusker Ojas KaleTaylor Futral tildabio/main:.github/CODEOWNERS /tusker/
virtual-assistant Sambasiva BandarupalliAhmed Aburaed tildabio/main:.github/CODEOWNERS /data-science/
workplan-agent Suresh BoddapatiMadhava Mukkamala tildabio/main:.github/CODEOWNERS /workplan_agent/
workplansvc Suresh BoddapatiMadhava Mukkamala tildabio/main:.github/CODEOWNERS /workplansvc/

Owner = the first person on the line, backup = the second; teams on a line are told but never own. A line that names nobody (the default *) sends it to the leads. Fix a wrong row in CODEOWNERS itself — X-ray re-reads it daily.