finds.dev← search

// the find

george0st/qgate-sln-mlrun

★ 302 · Python · Apache-2.0 · updated Aug 2025

MLRun/Iguazio/Nuclio quality gate solution. The solution checks a quality of MLRun implementation/delivery.

An external quality gate test suite for MLRun, the Iguazio MLOps platform. It runs end-to-end acceptance tests covering feature store ingestion, pipelines, model serving, and data sources — useful if you're evaluating MLRun before a production rollout or need independent verification that an upgrade didn't break your workflows.

The test matrix is thorough for the feature store layer specifically: it covers seven ingestion sources (DataFrame, CSV, Parquet, SQL, Kafka, HTTP) across create, ingest, and pipeline stages with a clear ✅/✔/❌ status per scenario. The version compatibility history going back to MLRun 1.3.0 is genuinely useful for teams stuck on older Iguazio on-prem deployments. Docker helper scripts for spinning up Redis, Kafka, MySQL, and Postgres locally make it easy to get the full target matrix running without cloud dependencies. The HTML/TXT report output is clean and shows pass/fail per test case at a glance.

The model monitoring and performance test sections (TS10, TS11) are entirely unimplemented — if those are your actual concerns, this doesn't help you. Coverage for XGBoost and DNN model builds and serving is also missing. The project is effectively a one-person effort (george0st), so MLRun 1.8 support was still marked as planned in early 2025 and may lag new MLRun releases. There's no CI that runs these against a live MLRun instance, so the test suite itself could drift out of sync with MLRun's actual behavior between manual validation runs.

View on GitHub →

// want more like this?

We dig through GitHub every week and send a few repos picked for what you actually care about — each with an honest take like this one.

Get finds in your inbox → Search again →