Looker Health Check & Technical Audit

Most Looker instances grow by accretion. Three years in, there are five measures called revenue, dashboards that take forty seconds to load, a BigQuery bill nobody can explain, and a sql_always_where that was supposed to be temporary. A health check is a short, fixed-scope engagement that tells you precisely where you stand and what to fix first, in a report you can hand to your own team or to us.

Scope

Performance review from System Activity. We analyze the History, Dashboard and PDT Event Log explores on your instance to find the slowest explores and dashboards, the queries that run most often, the PDTs that rebuild too frequently or fail silently, and the users and schedules generating the most warehouse load. Every finding is tied to a specific query pattern, not a general recommendation.

LookML model review. The project is read end to end and linted. We report on join fan-out and symmetric aggregate risks, duplicated and conflicting measures, missing primary keys, persist_for versus datagroup misuse, explores that should use aggregate awareness, Liquid that is fragile, and fields that no one has queried in a year and should be hidden or removed.

Security and access review. Roles, permission sets and model sets are reviewed against least privilege. We test row-level security (access_filter, sql_always_where, required_access_grants) for gaps, check user-attribute settings that allow users to edit their own values, review embed secrets and embed user permissions, and flag API keys belonging to departed staff.

BigQuery spend review (or Snowflake, Redshift, Databricks). For BigQuery we map Looker query patterns to slot usage and bytes scanned, check whether you are on the right edition or on-demand model, look at partition pruning and clustering from the Looker side, and quantify what caching, datagroups and aggregate tables would save. Other warehouses get the equivalent review against their billing model.

What you receive

• A written report, typically 20 to 30 pages, with findings ranked by impact and effort
• A prioritized remediation backlog in your tracker of choice, each item scoped in hours
• Annotated LookML diffs for the quick wins (we do not just describe them)
• A System Activity dashboard installed on your instance so you can keep watching the metrics we used
• A one-hour walkthrough with your team, and a follow-up call 30 days later

Pricing and timing

The health check is a fixed-price engagement, priced in three bands by instance size (number of projects, explores and active users), and takes two to three weeks from access to report. It requires an admin-level, read-only user on your instance and read access to your LookML repositories. Ask us for the current price bands when you get in touch.

Many clients use the audit as the first step before a larger project: a Looker (Google Cloud core) migration, an AI & Gemini in Looker rollout, or a performance rescue. Others run it annually as a governance checkpoint.

Vistelio is an independent Looker consultancy and is not affiliated with or endorsed by Google. Contact us to scope a health check.