flush and shutdown

SignaturesCopy Anchor Link

flowmark.flush(): Promise<void>
flowmark.shutdown(): Promise<void>

Available on Node.js and browser clients. The Next.js adapter exposes neither; it drains request-scoped delivery automatically with after().

flushCopy Anchor Link

Waits for pending delivery without closing the client. Rejects with FlowmarkDeliveryError if the drain reports delivery/validation errors or queue overflow. A successful flush does not prove business work succeeded.

try {
  await flowmark.flush()
} catch (error) {
  console.error("Flowmark could not flush telemetry", error)
}

Recorded failures are collected by the drain and cleared afterward. Do not interpret a later clean flush as proof that previously lost events arrived.

shutdownCopy Anchor Link

Stops accepting new workflows, drains pending events, and releases resources. Repeated shutdown calls return the same result. Starting another run afterward throws FlowmarkConfigError; existing markers cannot restart delivery.

try {
  await flowmark.shutdown()
} catch (error) {
  console.error("Flowmark shutdown reported delivery failures", error)
}

Place this at a job/process exit boundary, not after every function using a shared client. It does not call done() or fail() for open runs.

Browser limitsCopy Anchor Link

Browser flush() can be used before full-page navigation, but neither it nor keepalive guarantees delivery during unload or offline operation. Before the first browser workflow, flush is a no-op; shutdown still closes the client.

Read delivery and diagnostics for queue limits and error reporting without replacing your business result.