dash-observe 0.1.3

Data observability for Databricks — freshness, volume, and schema-change monitoring

DashObserve — Databricks Library

Part of the Dashlibs suite — Databricks libraries built for business users.

Notebook-native data observability for Databricks — no external service, no agent to deploy.

• Freshness monitoring — alert when a table's most recent timestamp value is older than expected
• Volume monitoring — alert on row-count bounds, or on deviation from a rolling historical baseline
• Schema-change detection — alert when columns are added, removed, or change type since the last run
• Next-update prediction — estimate when a table will next be refreshed, based on its observed update cadence
• Volume forecasting — project row counts over the next N days / weeks / months using a linear trend fitted to historical observations

All monitor results are appended to a Delta history table, which feeds baselines, schema-diff comparisons, and the forecasting models for future runs.

Installation

%pip install dash-observe

Quick Start

import dashobserve
dashobserve.launch()   # Opens interactive UI in your Databricks notebook

What it looks like

Python API

from dashobserve import MonitorConfig, run_monitors, run_forecast

# Monitor
cfg = MonitorConfig(
    table="catalog.schema.orders",
    freshness_column="updated_at", max_staleness_minutes=60,
    min_rows=1000, volume_tolerance_pct=20,
    track_schema=True,
)
report = run_monitors(cfg, history_table="catalog.schema.observe_history")
report.display()
print(report.summary())

# Forecast (requires history built up from prior monitor runs)
forecast = run_forecast(
    table="catalog.schema.orders",
    history_table="catalog.schema.observe_history",
    n_periods=4, period="weeks",
)
forecast.display()

Part of Dashlibs

Library	Purpose
dash-dq	Data Quality
dash-synthetic	Synthetic Data Generation
dash-observe	Data Observability (freshness, volume, schema)
dash-ml	ML Model Monitoring
dash-ingest	Data Ingestion
dash-gov	Data Governance
dash-relate	Ontology & Lineage for AI
dash-ui	Shared UI components (PyPI: dash-uis)

Quality & Contributing

• 33 unit tests, zero Spark dependency to run them — pytest tests/ -v (freshness/volume/schema-diff checks are pure Python and fully covered; only the Spark/Delta glue in runner.py needs a live cluster)
• Lint-clean (ruff check dashobserve/), PEP 561 typed (py.typed)
• Every change ships through a reviewed pull request; CI (lint → test on Python 3.9–3.12 → build) gates every PR and every release
• See CONTRIBUTING.md for dev setup, CHANGELOG.md for release history, SECURITY.md to report a vulnerability, and CODE_OF_CONDUCT.md

License

Apache 2.0