SlashPeak docs
Connect BigQuery with metadata-only access, run scans, read dollar-ranked findings, and fix them with the Cognis AI DBA. This is the customer guide for SlashPeak.
SlashPeak is a warehouse cost analytics platform. BigQuery is live today. Snowflake and Databricks are on the waitlist, with the same metadata-only rule.
This site is the customer manual. Use it to connect a project, understand every screen, pick a plan, and grant the smallest IAM set that still lets scans run.
Metadata only
The scanner never requests roles/bigquery.dataViewer on your tables. The one optional exception is a linked billing export, a single dataset of cost rows. Scans read INFORMATION_SCHEMA, storage sizes, reservations, and (if you grant them) lineage and audit APIs. Table contents stay in your project.
Start here
Create an account
Sign up, choose a plan or the trial, and run the first scan.
BigQuery setup
Console or gcloud. Custom IAM role, JSON key, then Test connection.
Findings
16 BigQuery rules plus 7 dbt rules, each with a dollar estimate.
Plans and billing
Trial, Starter, Growth, Pro and Enterprise, and every limit.
What a scan does
- You paste a service account JSON key.
- SlashPeak tests IAM on every GCP project that key can see.
- You pick the granted projects to include.
- A scan reads job history, table metadata and storage.
- Rules write findings. Savings are estimated when you open the finding, from current catalog data.
A scan never changes your warehouse. Findings are recommendations. On Growth and up, the Cognis AI DBA can make fixes for you, with a separate key, only inside the rules you set, and with an owner's approval.
Product map
| Area | What you get |
|---|---|
| Dashboard | Spend, forecast, commit, top offenders, efficiency trends |
| Anomalies | Unusual rises, with the cause and alerts |
| Usage | Cost by table, query, user and failures |
| Cost centers | Spend split by department or business unit, with monthly budgets |
| Findings | Ranked savings with SQL and attribution |
| Reservations | Slot use vs what you pay for (trial, Pro, Enterprise) |
| Data | Storage cost by project, dataset and table |
| Views | View and materialized-view catalog |
| Lineage | What feeds what, and what breaks if you delete a table |
| dbt | dbt model cost, graph and dbt rules |
| Airflow | Cost per Airflow DAG and task |
| Tags | Labels and Resource Manager tags |
| Audit logs | Scans, BigQuery audit events and AI actions (Growth and up) |
| Billing summary | Estimate vs your real Google bill |
Cognis AI
| Feature | What it does | Plans |
|---|---|---|
| Cognis chat | Ask about your spend in plain English | Starter and up |
| AI query optimizer | A cheaper rewrite of one expensive query | Starter and up |
| Cognis AI DBA | Fixes findings: BigQuery changes and GitHub pull requests | Growth and up |
| AI and your data | What is sent to which AI model | — |
Need help?
- Troubleshooting for IAM, APIs and empty scans
- FAQ and Pricing FAQ
- Contact if a permission probe disagrees with the Console