SolarflareDB
Walkthrough / Connected application data

From separate sources
to a connected answer.

Work through a customer-impact question using the actual Northstar sample, then change the source data and verify the result.

Open the example

Start with the question

Which active accounts have an Enterprise subscription to a product that depends on a degraded service?

This question has three constraints: the account is active, the relevant subscription uses an Enterprise plan, and the product on that subscription reaches a degraded service through its dependencies.

The important word is that. A healthy Enterprise subscription and an unrelated affected Team subscription should not make an account qualify.

Connect the four inputs

SourceWhat it contributes
Customer directoryEight accounts, eight contacts, and their ownership relationships.
Application catalogFour products, three plans, eight services, and the product and service dependencies.
Subscription ledgerNine relationships, each linking an account, a product, and a plan.
Service observationsThree incidents and their affected services.

The Enterprise-exposure query reads service status from the catalog and subsequent changes. The incident-tracing query additionally uses the observation records. Source names describe fictional input snapshots, not live connectors.

Follow one complete path

Copperline has an Enterprise subscription to Checkout. Checkout runs on the Checkout API, which depends on the Order processor, which depends on the degraded Payments service.

One matching dependency path
Copperline
  → subscription (plan: Enterprise)
  → Checkout
  → Checkout API
  → Order processor
  → Payments service (degraded)

Use the result inspector to follow the named participants and source references. The subscription’s plan is included in the evidence even though the traversal continues through the product.

Verify the initial result

At revision 4, the query returns Aster Labs, Copperline, and Fieldwork. Aster Labs matches through Checkout and Analytics. Copperline matches through Checkout. Fieldwork matches through Analytics.

Juniper Works also uses Checkout, but its subscription is on the Team plan. Meridian Systems has an Enterprise Identity subscription, but Identity is operational. Neither qualifies for this question.

Apply a change and rerun

In Data sources, apply Resolve the payment incident. The new observation sets Payments to operational and its incident to resolved in a single revision.

At revision 5, Copperline no longer qualifies. Aster Labs and Fieldwork remain because Reporting is still degraded and both have Enterprise Analytics subscriptions. The count changes from three to two because the stored records changed, not because a presentation step replaced a canned result.

Choose revision 4 and rerun to verify the earlier answer. The same query now reads the earlier snapshot, so it returns the original three accounts.

Change the model, not just the question

To explore beyond the sample, import the Asset CSV from the getting-started guide. Define an asset_owner relationship between Asset and Account, connect an asset to an account, and build a query that follows it.

The graph, record inspector, and query builder use the newly defined schema. There is no special-case result table for your imported data.

The sample demonstrates data modeling, bounded traversal, source attribution, and revision history in a local reference engine. It does not establish real-world incident impact or production service guarantees.
← All documentationBuild your own workspace →