Auditing Warehouse View Health

Auditing data warehouse view health

This skill produces a project-wide audit of materialized views (materialized saved queries) in the data warehouse — which ones are failing, and which are materialized but unused. Use it when the user wants a summary of view health, not a deep-dive on one failure.

The same underlying endpoint (data-warehouse-data-health-issues-retrieve) also reports source, sync, batch-export, and transformation issues. Source and sync health is covered by auditing-warehouse-source-health. Destinations (batch exports) and transformations are owned by other products — surface them if they appear, but route them to the relevant team rather than diagnosing here.

When to use this skill

• “Which of my views are broken?” / “Why is this materialized view failing?”
• “Are any of my materialized views wasting compute?”
• Reviewing view health after a HogQL or schema change
• Dashboards backed by materialized views are stale or erroring

Available tools

Filter the data-health-issues results to the materialized_view type for this audit. Use view-list when you need more than the active-failure summary (non-failing views, materialization flags, last-queried info) and view-run-history to see the run trail for a specific view.

What counts as a view “issue”

From the data-health endpoint, this audit cares about one of the five categories:

Each entry includes id, name, type, status, error, failed_at, and url.

The other categories the endpoint returns are out of scope for this skill:

• source / external_data_sync → auditing-warehouse-source-health
• destination (batch export) → owned by the batch exports / data pipelines product
• transformation (HogFunction) → owned by the CDP / ingestion side

Note the data-health endpoint only reports active failures. For views it doesn’t flag:

• Non-materialized views with errors (only materialized views are reported)
• Materialized views that are healthy but unused (costing compute every run) — see Step 4

Workflow

Step 1 — One-shot pull

Call data-warehouse-data-health-issues-retrieve and keep the materialized_view entries.

If there are no view issues, tell the user their materialized views are healthy and stop. Don’t invent problems.

Step 2 — Triage failures

Materialized view failures are usually independent of sources — a view failure is a HogQL or data issue in the view itself (syntax error, missing table reference, type mismatch). For each failing view, surface the error and point at the offending query. Use view-run-history if the user wants the failure trail.

Step 3 — Present the audit

Render a prioritized report. Don’t dump the raw JSON — human-readable:

## Materialized view health — 2 issues

### 🟠 Materialized views (2)
- monthly_revenue — view failed (syntax error in HogQL: 'FORM' instead of 'FROM')
- active_users_30d — view failed (missing table reference)

Both are HogQL issues in the view definitions — independent of your sources. Want me to open one?

Step 4 — Go beyond active failures (when asked)

Unused materialized views: Call view-list. Materialized views cost storage and compute every run. If any are marked materialized but haven’t been queried lately, surface them as cleanup candidates (the data is available via view-list; unmaterialize via view-unmaterialize).

Only run this extra check if the user explicitly asks for a broader audit.

Step 5 — Offer the next step

End the audit with a clear hand-off — e.g. “Want me to open monthly_revenue and fix the HogQL?” Never apply fixes autonomously from an audit; confirm explicitly before editing or unmaterializing a view.

Important notes

• The audit is read-only. Never call destructive tools (e.g. view-unmaterialize, view-delete) from the audit flow without explicit confirmation.
• Empty = healthy. Don’t pad an empty audit with hypothetical issues. “No view issues found” is a good answer.
• View failures are usually self-contained. Unlike source failures, a failed materialized view rarely cascades — it’s a query problem in that view. Don’t imply a broader outage.
• Sources, syncs, destinations, and transformations are out of scope here. They share the data-health endpoint but belong to other audits/products — route, don’t diagnose.