Datasetcredit-nov
Ava L.

rootplant-credit-nov-20261104.zip is imported. The file is a bundle, so this dataset has Bundle Analysis beside Traces & Spans.

credit-nov

This page shows what the dataset contains and what Auga found in it.

Source
rootplant-credit-nov-20261104.zip 1.9 MB
sha256
c25aafc5fb48…
Received
4 Nov 2026 14:44 UTC
Window
06:00 – 17:59 UTC (12.0 h)
Period
900 s
Analysis run
otlp-c25aafc5fb48
State
Active
Traces
463
399 flow · 64 sync
Spans
4,619
10.0 per run on average
Signals after screening
8of 14
6 set aside · example values until the API route exists
Findings
5
3 grade A · 1 grade C

Traffic over time

Runs per 15 minutes. The marks show where Auga found a link to one period later.

Traces per 15 minutes over the window0510152006:00Z08:00Z10:00Z12:00Z14:00Z16:00ZAuga: feed_batch_records → score_source changes here
runs per 15 minAuga's finding

Every run

The chart shows the duration of each run over time, with one dot per run.

Duration of every run over time0 ms2.5 s5.0 s7.5 s10.0 s
one run (small dot: a sync run with 2 spans)

Spans per run

The runs have 4 different lengths.

Runs by number of spans2 spans6410 spans7611 spans14512 spans178
Rootplant's credit template produces 10–12 spans in each flow run. The runs with 2 spans are sync runs (shared state). Check any run with a length outside the template's range.

Services

Auga computes the rate, errors and duration for each service from the spans.

ServiceSpansTracesp50p95Errors
bureau-agent894447916 ms1,993 ms0
income-agent798399553 ms853 ms0
intake-agent550399407 ms838 ms0
loan-orchestrator5133994,058 ms5,277 ms0
kyc-agent511411465 ms892 ms13
risk-agent499401450 ms866 ms0
policy-agent455401527 ms861 ms0
affordability-agent399399541 ms865 ms0

Who calls whom

The orchestrator, the agents in call order, and the tools each one uses. Dashed lines are the grade A links that Auga found.

Who calls whom: the orchestrator, the agents in call order, the tools each one runsloan-orchestratorintake-agent550 spans · p95 838 msDocument parsekyc-agent511 spans · p95 892 msSanctions screenSanctions list syncincome-agent798 spans · p95 853 msIncome verifybureau-agent894 spans · p95 1,993 msBureau fetchBureau feed syncaffordability-agent399 spans · p95 865 msrisk-agent499 spans · p95 866 msRisk agentScore model releasepolicy-agent455 spans · p95 861 msPolicy enginePolicy config deploy

Span names

19 distinct

Span nameSpansShare
Bureau agent invoke agent447
KYC agent invoke agent411
Risk agent invoke agent401
Policy agent invoke agent401
Loan orchestrator invoke agent399
Intake agent invoke agent399
Income agent invoke agent399
Income verify execute tool399
Bureau fetch execute tool399
Affordability agent invoke agent399
Document parse execute tool151
Notify applicant execute tool114
Risk agent chat96
Sanctions screen execute tool88
and 5 more

Signals

Each row shows one key in the file, at the boundary (edge.yaml) and after screening.

KeySpansBoundaryScreening
gen_ai.request.model3,256dropped · not_allowed—
http.route463dropped · not_allowed—
credit.feed_batch_records447allowkept
credit.policy_version401vocabconstant
credit.requested_amount399allowno link found
applicant.email399dropped · pii—
gen_ai.usage.input_tokens399dropped · not_allowed—
credit.applicant_segment399vocabno link found
credit.income_verified_ratio399allowkept
credit.document_text_layers399allowno link found
credit.score_source399vocabkept
credit.bureau_score399allowkept
credit.affordability_ratio399allowkept
credit.risk_score399allowkept
credit.decision399vocabkept
credit.feed_version48vocabconstant
credit.sanctions_list_version12vocabkept
credit.model_version2vocabconstant
A dropped key follows the policy and is not a problem, so it has the normal text colour. Red marks only a key that the template promised and that is missing from the file.

Enough data?

The chart compares the number of periods at each length with the rule of at least 30, and marks the suggested period.

Periods by length60 s720900 s ·481500 s29≥ 30
enough periodsnot enough (striped)the rule of at least 30
60 s: 720 periods, 51% empty · 900 s: 48 periods, 0% empty · 1500 s: 29 periods, 0% empty. Auga's rule: discovery needs at least 30 periods left after screening has used its share.

Screening

Of 14 candidates, screening kept 8. The values are examples until the API route GET …/summary exists.

Signals by screening resultkept8no link found3constant3
A constant signal never changes, so it cannot have a link. A duplicate is the same signal under two keys. "No link found" means that nothing changed together with the signal inside the time window.

Links found

What Auga found in run otlp-c25aafc5fb48. Grade A means identified, and grade C means a guess.

CauseEffectGradeLagStrengthIn words
credit.feed_batch_recordscredit.score_sourceA10.93An empty feed batch makes the bureau score stale one period later. We are very sure: 3 of 3 challenge tests held.Prediction If the feed batch is empty again, stale scores follow within one period. · recheck 11 Nov 2026
credit.score_sourcecredit.risk_scoreA00.91A stale bureau score raises the risk score in the same period. We are very sure: the effect held in 6 other environments.Prediction A stale score will raise the risk score by about 12 points. · recheck 11 Nov 2026
credit.risk_scorecredit.decisionA00.96A higher risk score makes a denial more likely in the same period. We are very sure: 4 of 4 registered predictions held.Prediction Above a risk score of 62, denials increase. The next 30 decisions will show it. · recheck 8 Nov 2026
credit.income_verified_ratiocredit.affordability_ratioB00.78We are fairly sure that a better-verified income is followed by a higher affordability ratio. One of three challenge tests did not hold.no registered prediction · recheck 18 Nov 2026
credit.sanctions_list_versioncredit.decisionC20.41Auga has no plain wording for this link yet (coincident). Ask for a wording review. · 422 ·

Traces

463 of 463. Open shows each step of a run in time order. A click on a dot, bar or link above brings you here.

TimeRoot spanSpansDurationStatus
06:00ZBureau agent invoke agent22,107 msokOpen
06:02ZLoan orchestrator invoke agent125,832 mserrorOpen
06:04ZLoan orchestrator invoke agent123,248 mserrorOpen
06:05ZLoan orchestrator invoke agent103,945 msokOpen
06:06ZLoan orchestrator invoke agent114,726 msokOpen
06:07ZLoan orchestrator invoke agent104,219 msokOpen
06:09ZLoan orchestrator invoke agent104,846 msokOpen
06:12ZLoan orchestrator invoke agent123,693 msokOpen
06:13ZLoan orchestrator invoke agent104,425 msokOpen
06:15ZBureau agent invoke agent22,063 msokOpen
06:15ZKYC agent invoke agent21,289 msokOpen
06:15ZLoan orchestrator invoke agent124,493 msokOpen
At most 50 of 463 are listed. Use the search to see the rest.

Example data: this page generated a Rootplant-style credit dataset from seed 20261104 (463 runs, 4,619 spans). Auga's figures for screening and findings are example values until that part of the product is built. This is a static example. The search fields do not work, and every Open link leads to the same example run.

What p50 and p95 mean

Half of the spans of a service were faster than p50, and 95% of them were faster than p95.

More exactly

p50 is the median span duration for each service. p95 is the 95th percentile.