Getting Started

This page walks through the path from "we just signed up" to "we're seeing recommendations land in our inbox." Most teams complete it in under 30 minutes.

What QueryWise needs to be useful

QueryWise needs three categories of data, in roughly this order:

1. Billing data from each cloud / database vendor. This is what powers cost views.
2. Workload metadata — query stats, explain plans, schema, usage metrics. This is what powers recommendations.
3. Tags / cost centers — optional but high-leverage. This is what powers chargeback and unit economics.

You can connect (1) without giving QueryWise database credentials at all. (2) requires read access to the workload data; we never write or modify customer data. (3) is configured inside QueryWise once tags are flowing.

Step 1: Sign up and create a tenant

Sign in via your identity provider (Auth0 with Google, Microsoft, or email/password supported). On first login you'll be asked to name your tenant — typically your company name. The tenant is the isolation boundary for everything you do in QueryWise.

The first user automatically gets the owner role. Roles, in decreasing privilege:

• owner — full control, billing, team management
• admin — full operational control
• editor — connect accounts, apply recommendations
• viewer — read-only

You can invite more users from Settings → Team.

Step 2: Connect your first cloud account

Go to Settings → Cloud Accounts and click "Add account." Pick the vendor.

The simplest path is IAM-only / API-key onboarding — no database credentials required:

• AWS — Cross-account IAM role via CloudFormation Quick Create. Click the button, deploy the stack, paste the role ARN.
• GCP — OAuth + service account, with optional cloud-shell-driven setup.
• Azure — Managed Identity for resource and cost APIs.
• Snowflake — Username + password or RSA key-pair (recommended).
• Databricks — Personal Access Token (PAT).
• MongoDB Atlas — Programmatic API key.

For the full credential matrix and least-privilege IAM policies, see Cloud Connectors.

After saving, QueryWise immediately runs a connection test. The first sync starts in the background.

Step 3: Wait for the first sync

The first sync typically takes 5–20 minutes depending on history depth and account size. You can watch progress on the cloud-account row.

Two things happen during the first sync:

• Billing data is backfilled (default: 90 days).
• Workload metadata sync starts on a schedule (every 15 min for query stats, daily for schema metadata).

Once the first sync completes, the dashboards start populating.

Step 4: Tour the dashboards

Five pages are worth opening on day one:

• Overview — KPI summary, spend trend, top cost drivers
• Costs — vendor / service / pipeline breakdowns, AI-specific tab
• Queries — top queries by cost, with explain plan + anti-pattern panels
• Recommendations — what QueryWise thinks you should fix, with savings estimates
• Allocation — tag-based cost allocation; later, chargeback

You don't need to do anything to populate these — once data is flowing, they fill themselves.

Step 5: Set up tags and a cost center

Open Allocation → Tag Discovery to see which tags are present on your cloud accounts and how much spend each covers. The first goal is honest visibility, not perfection.

If you want chargeback: define one or two cost centers in Settings → Cost Centers. Even a rough split (engineering / data / product) is more useful than one bucket.

For the full tagging workflow, see the Cost Attribution Foundations guide.

Step 6: Apply your first recommendation

By the end of the first sync, you'll usually have 10–50 recommendations. Sort by estimated savings.

The categories that are safest to apply first:

• Auto-suspend tweaks on Snowflake warehouses
• Idle endpoint shutdowns on AWS / Azure
• Unused index removal on RDS / Cloud SQL / Azure SQL

Each recommendation has a remediation snippet, an estimated savings figure, and an apply path (manual, ticket sync to JIRA/ServiceNow, or auto-apply if you've defined a policy).

What to do next

• Connect more cloud accounts. The platform gets sharper with more data.
• Read the Query Optimization Playbook and start tuning.
• Set up Slack or email alerts in Settings → Notifications so anomalies don't sit in a dashboard waiting for you.

If anything is unclear or sync isn't progressing, the /assistant page has an AI agent that can investigate connection issues, sync state, and recommend next steps.