> ## Documentation Index
> Fetch the complete documentation index at: https://docs.planasonix.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Cost insights

> Track and optimize data platform costs.

**Cost insights** attribute warehouse spend, compute minutes, and connector activity to pipelines and teams. You answer “what drove last month’s bill?” with evidence instead of guesses.

## Cost tracking

Costs roll up from:

* **Warehouse queries** executed by Planasonix on your behalf (when telemetry is available from the warehouse)
* **Managed compute** for Planasonix-hosted runners or integrated clusters
* **Egress and API** charges where cloud billing APIs map usage to jobs

Tag pipelines with `team`, `product`, and `env` so finance can allocate spend fairly.

<Steps>
  <Step title="Connect billing sources">
    Admins link cloud billing exports or warehouse usage APIs in **Settings** → **Compute** / integrations (per deployment).
  </Step>

  <Step title="Validate attribution">
    Run a known test job and confirm its cost appears under the expected pipeline within 24–48 hours.
  </Step>

  <Step title="Review monthly">
    Compare top pipelines week over week; investigate new entrants after releases.
  </Step>
</Steps>

## Usage analytics

Usage charts show **CPU-hours**, **slot time**, **bytes scanned**, or **rows processed** depending on the engine. Slice by schedule vs ad-hoc runs to see whether automation or manual previews dominate.

<Tabs>
  <Tab title="Engineers">
    Find expensive nodes or queries to refactor first (wide scans before selective filters).
  </Tab>

  <Tab title="Managers">
    Compare teams’ relative usage against roadmap priorities.
  </Tab>
</Tabs>

## Optimization recommendations

Recommendations may suggest:

* Cluster rightsizing for recurring jobs
* Partition pruning opportunities in SQL
* Caching or incremental patterns instead of full reloads
* Moving dev workloads to smaller warehouses

Treat each suggestion as a hypothesis—validate with your warehouse’s query profiler before changing SLAs.

<Info>
  Some warehouses bill separately for storage; cost insights may focus on compute unless you integrate storage billing feeds.
</Info>

## Related topics

<CardGroup cols={2}>
  <Card title="Compute settings" icon="microchip" href="/settings/compute">
    Configure execution environments that drive spend.
  </Card>

  <Card title="Usage and limits" icon="chart-column" href="/settings/usage">
    Compare actuals to plan entitlements.
  </Card>
</CardGroup>