@spec cancel(GenServer.server()) :: :ok

Cancels the given scheduled job

@spec run_at(Job.task(), DateTime.t(), keyword()) :: GenServer.on_start()

Runs the given function at the given time

@spec run_at(module(), atom(), list(), DateTime.t(), keyword()) ::
  GenServer.on_start()

Runs the given module, function and argument at the given time

@spec run_every(Job.task(), String.t() | Crontab.CronExpression.t(), keyword()) ::
  GenServer.on_start() | {:error, any()}

Runs the given function on every occurrence of the given crontab. If func is of arity 1, the scheduled execution time will be passed for each invocation

Takes the same options as run_every/5

@spec run_every(
  module(),
  atom(),
  list(),
  String.t() | Crontab.CronExpression.t(),
  keyword()
) ::
  GenServer.on_start()

Runs the given module, function and argument on every occurrence of the given crontab. Any values in the arguments array which are equal to the magic symbol :sched_ex_scheduled_time are replaced with the scheduled execution time for each invocation

Supports the following options:

• timezone: A string timezone identifier (America/Chicago) specifying the timezone within which the crontab should be interpreted. If not specified, defaults to UTC
• time_scale: A module that implements the Fledex.Scheduler.TimeScale behaviour, by default is set to Fledex.Scheduler.IdentityTimeScale. Can be used to speed up time (often used for speeding up test runs)
• name: To attach a name to the process. Useful for adding a name to Registry to lookup later. ie. {:via, Registry, {YourRegistryName, "scheduled-task-1"}}
• nonexistent_time_strategy: How to handle scheduled runs within a DST forward boundary when evaluated within the timezone specified by the timezone option. Valid values are :skip (the default) and :adjust. By way of example, for a job which is scheduled to happen daily at 2:30am in the America/Chicago timezone, on days where a forward DST transition happens (such as 10 March 2019) :skip will skip this invocation and next run the job at 2:30 CDT 11 March 2019, while :adjust will run the job the same amount of time into the day as it would normally run (2.5 hours after midnight, which will be at 3:30 CDT 10 March 2019).

@spec run_in(Job.task(), Job.schedule() | pos_integer(), keyword()) ::
  GenServer.on_start()

@spec run_in(Job.task(), pos_integer(), keyword()) :: GenServer.on_start()

Runs the given function in given number of units (this corresponds to milliseconds unless a custom time_scale is specified). If func is of arity 1, the scheduled execution time will be passed for each invocation

Takes the same options as run_in/5

@spec run_in(module(), atom(), list(), pos_integer(), keyword()) ::
  GenServer.on_start()

Runs the given module, function and argument in given number of units (this corresponds to milliseconds unless a custom time_scale is specified). Any values in the arguments array which are equal to the magic symbol :sched_ex_scheduled_time are replaced with the scheduled execution time for each invocation

Supports the following options:

• repeat: Whether or not this job should be recurring
• start_time: A DateTime to use as the basis to offset from
• time_scale: A module that implements the Fledex.Scheduler.TimeScale behaviour, by default is set to Fledex.Scheduler.IdentityTimeScale. Can be used to speed up time (often used for speeding up test runs)
• name: To attach a name to the process. Useful for adding a name to Registry to lookup later. ie. {:via, Registry, {YourRegistryName, "scheduled-task-1"}}

@spec stats(GenServer.server()) :: Fledex.Scheduler.Stats.t()

Returns stats on the given job. Stats are returned for:

• scheduling_delay: The delay between when the job was scheduled to execute, and the time it actually was executed. Based on the quantized scheduled start, and so does not include quantization error. Value specified in microseconds.
• quantization_error: Erlang is only capable of scheduling future calls with millisecond precision, so there is some inevitable precision lost between when the job would be scheduled in a perfect world, and how well Erlang is able to schedule the job (ie: to the closest millisecond). This error value captures that difference. Value specified in microseconds.
• execution_time: The amount of time the job spent executing. Value specified in microseconds.