Python lifecycle

asyncio maker groups, explicit construction phases, source-last start, and deadline-aware shutdown.

When to use

Use this page when implementing maker hooks, owning resources, or diagnosing startup and shutdown.

Behavior

custom_makers_init runs once before infrastructure and function creation.
_MakerGroup uses asyncio.gather, cancels its context, and raises the first error.
build_stream_graph precedes bind_transports; managed connector cores start before graph resources, then Worker admission and ordinary Sources open before the aiohttp listener.
stop_service bounds Context once, calls on_stop, then runtime stop with the remaining deadline.

Python startup sequence

ConfigurationLoad graph configResolve typed service, stream, connector, and endpoint settings.

Factories_initialize_makersInstall defaults, then run custom_makers_init, custom_functions_init, and on_start.

ConstructionInvoke makers onceCreate infrastructure and long-lived functions in parallel.

BindingBuild graph and endpointsBind streams, callers, endpoint consumers, functions, and transport routes.

Connector coreOpen clientsStart managed connector clients before anything can submit or poll.

DownstreamStart graph resourcesStart storage, pools, components, and Sinks before inbound work.

Worker admissionStart managed pollingOpen Temporal or other connector-owned inbound workers after Sinks are ready.

SourcesStart DataSourcesOpen ordinary graph admission only after every downstream dependency is ready.

ServersStart listenersOpen generated HTTP and gRPC listeners after endpoint bindings and graph resources are ready.

Python shutdown sequence

SignalStop requestedCreate one deadline from shutdownTimeout.

NetworkClose listenersReject new HTTP/gRPC requests and drain accepted handlers while graph dependencies remain alive.

SourcesDrain root workStop ordinary DataSources and admission-aware components.

WorkersStop managed admissionStop connector-owned polling only after ordinary Sources have drained.

ProducersPools, timers, storageStop admission and drain tasks that can still create graph work.

GraphParallelCall registryDrain only after no producer can create another child.

OutputStop DataSinksFlush results and accepted outbound submissions.

Connector coreClose clientsRelease managed connector clients after Sources, graph work, and Sinks no longer need them.

ReleaseRelease graph ownershipDrop endpoint consumers, streams, functions, and bindings only after their callbacks cannot run.

TelemetryFlush diagnosticsMetrics, traces, and logs cover shutdown and stop at the final safe boundary.

Config reload, admission-aware Components, and aiohttp cleanup close first; accepted HTTP handlers retain endpoint state and graph dependencies while draining.
Ordinary DataSources and regular Components stop next.
Managed connector Worker admission stops only after ordinary Sources drain.
DelayPool, TaskPools, PriorityTaskPools, deferred Components, and storage drain before the service task registry.
After producers quiesce, registry snapshots drain repeatedly because one ParallelCall may create another.
DataSinks stop before managed connector clients; metrics and tracing stop before logs.
Timed-out asyncio shutdown operations are cancelled and observed.

Hooks and boundaries properties

Generated extension points and runtime-owned lifecycle boundaries.

PropertyTypeDescription

custom_makers_initasync hook

Mutates generated makers once.

initialize_infrastructuregenerated

Creates servers, clients, and infrastructure.

initialize_functionsgenerated

Runs function makers concurrently.

on_start / on_stopasync hooks

Run around runtime activation and drain.

run_shutdown_operationsruntime

Uses only the remaining shared deadline.

Lifecycle outline

configure -> make once -> bind -> start -> admit -> drain -> stop