Struct Worker 

pub struct Worker { /* private fields */ }

Any number of workers can share a database; each message is handled by one.

Implementations

impl Worker

pub fn job<Args: 'static>( self, name: &'static str, job: impl Call<JobContext, Args>, ) -> Self

impl Worker

pub fn new(db: Db) -> Self

pub fn on<E: Event, Args: 'static>( self, name: &'static str, job: impl Call<JobContext, Args>, ) -> Self

Runs job for each E raised.

pub fn every<Args: 'static, J: Call<JobContext, Args>>( self, name: &'static str, cadence: impl Into<Cadence>, job: J, ) -> Self

At an interval, first as soon as the schedule is new; or at a WallClock time, first when the zone’s clocks next show it.

pub fn each<A: Aggregate, QueryArgs: 'static, Args: 'static>( self, name: &'static str, interval: Duration, items: impl Call<JobContext, QueryArgs, Outcome = Vec<A>>, action: &'static str, job: impl Call<JobContext, Args>, ) -> Self

Each record’s job runs with its own retries.

pub fn mirror<M: Mirror>(self, name: &'static str) -> Self

A new mirror first copies every record its source has, once.

pub fn max_attempts(self, max_attempts: NonZeroU16) -> Self

At least one, as none would leave every message unhandled:

fn never(worker: sofie_db::worker::Worker) -> sofie_db::worker::Worker {
    worker.max_attempts(0)
}

pub fn batch(self, size: NonZeroUsize) -> Self

pub async fn run(self, shutdown: impl Future<Output = ()> + Send) -> Result<()>

Refuses repeated job names, intervals out of range and an unkept feed before any work.

pub async fn drain(&self) -> Result<usize>

Pulls every feed first; a failed job is not counted, it is retried once its backoff is over.

pub async fn drain_reporting(&self) -> Result<Drained>