Bosca / Gateway

Upstreams

A service the proxy knows how to reach

Each upstream is one URL plus everything about talking to it well — a health probe, timeouts, and a connection pool — managed as a record in Studio, not a config file on a box.

Registration

The address, and how to treat it

Point the gateway at the service and set how the proxy should treat it — per upstream, because a query engine and a tiny dashboard don't deserve the same timeouts.

  • The address — an HTTP or HTTPS URL, typically an internal service with no public address of its own.
  • A health probe — a path the proxy checks on the interval you choose.
  • Timeouts — connect and request limits, per upstream.
  • A connection pool — how many idle connections to keep, and for how long.
upstream · warehouse
URLhttp://warehouse:8080
Health check/v1/info · every 30s
Request timeout60s
Pool8 idle · 90s

Health

You'll know before your users do

The proxy probes each upstream on its interval and reports what it finds into Studio — the status, when it last changed, and the reason when something failed.

  • Up — the last probe on the configured path succeeded.
  • Down — the probe failed; the failure reason is shown on the upstream's page.
  • Unknown — no probe configured, or none has run yet.
upstream health
warehouseUP · 30s probe
dashboardsUP · 60s probe
legacy-reportsDOWN · connection refused

The switch

On, off, and accounted for

Routes and upstreams each carry their own enabled switch, and the relationships between them stay visible.

  • Disable a route and it stops matching — the upstream and its other routes are untouched.
  • Routes are listed where you need them — the upstream's page shows every route bound to it, with its pattern, sign-in method, and groups.
  • Deleting is deliberate — an upstream's routes are removed first, then the upstream.
routes on warehouse
/warehouse/**OAuth2 · analysts
/warehouse-api/**JWT · data-eng

Keep exploring

More on Gateway