Production benchmarks

Run the pipeline.
Keep the data.

These aren't synthetic benchmarks. They're three real pipelines we ran on SentraZero — retail comparison, image embedding, and duplicate detection — all executed on agents inside our own infrastructure. The control plane never saw a single byte of our data. Here's the proof.

100%
Pipeline success rate
0
Client bytes in control plane
<90s
Wall time, every run
3
Active agents · ARM + amd64
Execution log

Every step,
every millisecond.

SentraZero logs every plugin execution, every agent assignment, and every timing metric. Here's what three real production runs actually looked like — straight from the agent, never from the control plane.

Pipeline execution

$ sentra job run --pipeline walmart-scrape-compare

Compound mode │ 9 URL pairs │ scrape→compare→merge │ ~65s

✓ Completed — 100% success

$ sentra job run --pipeline cat-image-embed

Compound mode │ 202 images │ load→resize→embed │ 27–81s

✓ Completed — 100% success

$ sentra job run --pipeline walmart-duplicate-detection

Compound mode │ 19 products │ search→classify │ 14.4s

✓ Matched ground-truth 19/19 (100% correct)

$ _

The Pipelines

One engine. Every data type.

Structured rows, images, and product catalogs — three completely different workloads, all orchestrated by the same self-hosted engine. That's the point: you bring the data, SentraZero brings the pipeline.

Walmart Scrape & Compare

Retail catalog validation

Compares products across retailers by scraping live pages and matching them against a reference dataset. Catches price discrepancies and listing errors in under a minute, with every decision logged for audit.

Dataset9 URL pairs
StepsScrape → Compare → Merge
Wall time~65s (compound)
Success100%

Cat Image Pipeline

ML data preparation

Loads 202 images, resizes them, and produces 809-dimensional hand-crafted numpy embeddings — color histogram, edge, texture, and channel stats. Output lands in the client's own object storage, never ours.

Dataset202 images
StepsLoad → Resize → Embed
Embedding809-dim numpy
Wall time27–81s

Duplicate Detection

Catalog reconciliation

Searches and classifies 19 products, then reconciles against the client's own ground-truth Baselining.xlsx. Matched the reference exactly — 19 of 19, 100% correct — in 14.4 seconds.

Dataset19 products
StepsSearch → Classify
Wall time14.4s
Accuracy19/19 matched
Pipeline 1

Walmart Scrape & Compare

This pipeline answers a question every retail and ops team asks: “Do our listings actually match what's on the marketplace?” It scrapes live product pages, compares them against a reference dataset, and flags every discrepancy — all on your infrastructure, with raw pages never touching our servers.

Why this matters

A single mismatched price erodes margin and trust. SentraZero automates the whole loop — scrape, compare, merge — in about 65 seconds of compound-mode wall time, and logs every verdict so it's audit-ready.

Honest about limits

Some matches were capped only because Walmart geo-blocked outbound requests — an external source constraint, not a platform defect. The engine did its job; the wall did its job too.

0
scrape
compound

9 URL pairs fetched from retailers. Playwright handles JavaScript-rendered pages; a lightweight fallback covers the rest. Automatic retry on failure keeps the run green.

1
compare
compound

Rows compared using similarity on product titles and brands. Exact-match, brand-similarity, and title-similarity thresholds plus a UPC fast-path keep matching precise and fast.

2
merge
compound

Results merged back into the original dataset. Each row carries a decision (match / no_match / uncertain), a confidence score, and the reasoning behind the verdict.

Plugin Configuration

Compare methodSimilarity
Exact threshold0.6
Brand threshold0.8
Title threshold0.8
UPC bypassEnabled
ModeCompound (~7× faster)
Pipeline 2

Cat Image Pipeline

This is what data sovereignty looks like in practice. 202 images are loaded, resized, and turned into 809-dimensional hand-crafted numpy embeddings — built from color histograms, edge, texture, and channel statistics — and the results are written straight to the client's own object storage. The control plane stores none of it.

Why this matters

Most platforms pull your imagery into their cloud to preprocess it. SentraZero flips that: the work happens where the data already lives. Only lightweight feature vectors leave the machine — raw images never do.

Integration Details

S3 accessClient object storage
BackendS3HTTP + SigV4
ExecutionLocal agent (arm64)
Output809-dim numpy vectors
0
image_load
fastest I/O

202 JPEG images pulled from the client&apos;s own object storage over a custom S3HTTP backend with AWS SigV4 signing. Cross-project access validated; nothing is copied to SentraZero.

1
image_resize
fast

All 202 images resized in a single batched pass, ready for downstream model input. Size-based batching keeps memory flat regardless of dataset size.

2
image_embed
fast

Each image becomes an 809-dimensional hand-crafted numpy embedding &mdash; color histogram, edge, texture, and per-channel statistics &mdash; written back to the client&apos;s storage. No raw pixels leave their perimeter.

Head to Head

Three workloads, one engine

Structured rows, images, and catalog reconciliation — different data, same orchestration. That's the “One engine. Every data type.” promise, measured.

MetricScrape & CompareImage EmbedDuplicate Detect
Data typeStructured rowsImagesCatalog products
Dataset9 URL pairs202 images19 products
StepsScrape → Compare → MergeLoad → Resize → EmbedSearch → Classify
Wall time~65s27–81s14.4s
Success100%100%19/19 matched
Output storedClient sideClient object storageClient side
Control-plane bytes000

Key insight

Compound mode ran roughly 7× faster than per-step execution — the engine batches the I/O-bound work instead of paying for it twice. The bottleneck in every pipeline is network I/O, not compute, which means adding agents parallelizes the slow part. Your data stays put; the work comes to it.

Engineering Fixes

Battle-tested on real runs

Real pipelines surface real edge cases. Three agent-binary fixes came directly out of these runs — each validated by re-running the pipeline until it was green. That's how a self-hosted platform earns your trust: it breaks in the open, then it doesn't.

01

Separated stdout / stderr in the sandbox

internal/plugin/sandbox.go

Problem

Plugin stdout and stderr shared one buffer. A stray warning turned valid JSON into garbage and a step would fail silently.

Fix

Split stdout and stderr into distinct OS pipes. Only stdout is parsed as JSON; stderr is captured for logs.

Impact

Reliability win for every plugin that emits warnings or framework logs. No more silent step failures.

02

Removed the 4 KB output ceiling

internal/plugin/sandbox.go

Problem

A hard 4096-byte cap truncated large plugin output mid-object, silently dropping records from the result.

Fix

Raised the output buffer from 4 KB to 10 MB so plugins can return full datasets with complete metadata.

Impact

Pipelines that process hundreds of records no longer lose data. The 202-image run depends on this.

03

Defensive JSON parser

internal/dispatcher/handlers_unix.go

Problem

Debug banners or warnings ahead of a JSON payload crashed json.Unmarshal at the parse site.

Fix

Added parsePluginOutputAsJSON() &mdash; a backward-scanning parser that finds the last valid JSON object. Rolled out to all 5 parse sites.

Impact

The agent now tolerates imperfect plugin output. Plugins can log freely without breaking the run.

Key Takeaways

Zero client data in the control plane.

The headline isn't speed — it's sovereignty. We verified it: the control-plane tables that would tempt a vendor to hoard your data held exactly zero rows.

01

Sovereignty, verified — not claimed

Control-plane tables vector_store, step_outputs, and plugin_execution_history contained 0 rows of client data. We checked. Your raw data never transits or rests in our systems.

02

One engine, every data type

CSV rows, 202 images, and 19-product catalog reconciliation all ran through the same orchestration. Structured, visual, and document data — no separate stack to buy or learn.

03

Signed, sandboxed, isolated

Plugins are Ed25519-signed and execute in a sandbox; orgs are isolated by Row-Level Security. Unauthorized code can't run, and one tenant can't see another's data.

04

Honest about what we can't control

When an external source (Walmart) geo-blocked requests, matches were capped — a wall on their side, not a defect in ours. We tell you that instead of hiding it.

100% success
3 real pipelines
0 bytes
in control plane
<90s
wall time, every run
3 agents
ARM + amd64

See the benchmarks for yourself

Deploy an agent on your own infrastructure. Point it at your sources. Watch real pipelines run with every step logged, every byte staying home, and every metric auditable. See the verified benchmarks for yourself.