Ringly (Postgres)

Ringly is a made-up SaaS company selling a cloud phone system (calls and SMS) to support and sales teams. This folder is a complete fastdata setup for it: a 15-table Postgres database, a read-only role, a connector config and a knowledge base.

The data is small on purpose: 6–7 rows per table, consistent across tables, so you can check every answer by hand.

Tables#

Table and column comments are in the schema (initdb/01-schema.sql), and the knowledge base in knowledge/ describes each table, the business terms (MRR, answer rate, NPS) and verified example queries.

Run it#

With Docker (Postgres 17 on localhost:5433, with the real read-only role):

cd examples/voicecall
./seed.sh                 # starts Postgres and loads the data; run again to reset
cp .env.example .env
fastdata doctor
fastdata

Without Docker (Postgres compiled to WASM, in memory; from the repository root):

npm install
npm run example:ringly    # leave it running
# in another terminal:
cd examples/voicecall
RINGLY_DB_URL=postgres://fastdata_ro:fastdata_ro@localhost:5433/postgres fastdata

PGlite does not enforce roles, so use Docker to test the read-only role itself.

Your own Postgres: create an empty database named ringly, then DATABASE_URL=postgres://owner@host/db ./seed.sh --url (needs psql), and point RINGLY_DB_URL in .env at the fastdata_ro role.

Questions to try#

• What is our current MRR by plan?
• Which customers have the worst answer rate on inbound calls?
• Which customers are at risk of churning, and why?
• What is our NPS, and what do detractors complain about?
• Which invoices are overdue or still open?
• Which integrations are failing, and does it show in support tickets?

Some answers, for checking: MRR is $4,906 (ARR $58,872); NPS is 0 (2 promoters, 2 passives, 2 detractors); Nordwind Energie has a failing Zendesk integration, an urgent ticket about it and the lowest NPS score.

Files#