Changelog

What we've shipped lately. Most-recent first.

Improvement

Reliability sprint

  • Idempotent connection creation in onboarding — re-running the test step no longer creates duplicate connections.
  • Heatmap dashboard now reads from the canonical Tier 1 discovery output, eliminating duplicate table rows when validation rules used mixed naming.
  • Validation issue automation now gates on rule source — auto-generated statistical rules stay silent until reviewed, no more 20-issue blasts on a fresh signup.
  • Notification emails suppressed during onboarding (until users have configured preferences). Trial signups no longer receive validation-failure emails before reaching the dashboard.
Feature

14-day free trial

  • Launched the 14-day free trial: full product, all features, no credit card required.
  • Trial banner shows remaining days and a contact CTA when fewer than 2 days remain.
  • Trial expiration auto-downgrades to a limited Free plan; trial state is exposed via /api/billing/trial-status.
Feature

BigQuery onboarding

  • End-to-end BigQuery onboarding: connect with a service account, auto-discover datasets and tables, generate recommended validations.
  • View row-counts are enriched on the first 50 sampled views via bounded COUNT(*) queries.
  • DATE-aware freshness SQL — DATE columns get DATE_SUB(CURRENT_DATE()), DATETIME and TIMESTAMP get their matching functions.
Fix

Dashboard analytics retry hardening

  • All dashboard analytics routes (health-trend, issues-summary, top-failing-validations, connection-heatmap) now retry on Supabase HTTP/2 idle disconnects.
  • Top Failing Validations widget now queries the actual validation_results table instead of the higher-level issues tracker — surfaces all failing rules, not just escalated ones.
  • Data Objects Tracked card and connection-heatmap stuck-loading states resolved.
Feature

Anomaly detection (z-score, IQR, moving average)

  • Three anomaly detection algorithms available per profiled column: z-score, IQR, and moving average.
  • Anomaly history view with severity classification (low / medium / high) and acknowledgment tracking.
  • Per-connection anomaly scan automation; runs on the same schedule as Tier 4 profiling.
Feature

Smart alert routing and snoozing

  • Per-user notification preferences: receive failures as immediate alerts or daily digests.
  • Snooze rules per validation, table, or connection — useful for known-bad windows during deploys or backfills.
  • Multi-channel routing: Slack, email, with per-rule channel selection.
Platform

Tiered metadata collection

  • Five-tier metadata collection pipeline (Discovery, Schema, Changes, Profiling, Lineage) running on configurable schedules per connection.
  • Schema change detection across renames, type changes, adds, and drops — not just adds.
  • Auto-cached validation rule generation from Tier 4 profile statistics.

Want to try the latest?

14-day free trial, no credit card. Everything in the changelog is available in the trial.

Start Free Trial