luigi server config explained

luigi_server.cfg

[core]
# These parameters control core luigi behavior, such as error e-mails and
# interactions between the worker and scheduler.

default-scheduler-host: localhost
# Hostname of the machine running the scheduler. Defaults to localhost.

default-scheduler-port: 8082
# Port of the remote scheduler api process. Defaults to 8082.

# logging_conf_file:
# Location of the logging configuration file.

no_configure_logging: false
# If true, logging is not configured. Defaults to false.

parallel-scheduling: false
# If true, the scheduler will compute complete functions of tasks in
# parallel using multiprocessing. This can significantly speed up
# scheduling, but requires that all tasks can be pickled.

retry-external-tasks: false
# If true, incomplete external tasks (i.e. tasks where the `run()` method is
# NotImplemented) will be retested for completion while Luigi is running.
# This means that if external dependencies are satisfied after a workflow has
# started, any tasks dependent on that resource will be eligible for running.
# Note: Every time the task remains incomplete, it will count as FAILED, so
# normal retry logic applies (see: `disable-num-failures` and `retry-delay`).
# This setting works best with `worker-keep-alive: true`.
# If false, external tasks will only be evaluated when Luigi is first invoked.
# In this case, Luigi will not check whether external dependencies are
# satisfied  while a workflow is in progress, so dependent tasks will remain
# PENDING until the workflow is reinvoked.
# Defaults to false for backwards compatibility.

rpc-connect-timeout: 10.0
# Number of seconds to wait before timing out when making an API call.
# Defaults to 10.0

# smtp_host:
# Hostname for sending mail throug smtp. Defaults to localhost.

# smtp_local_hostname:
# If specified, overrides the FQDN of localhost in the HELO/EHLO
# command.

# smtp_login:
# Username to log in to your smtp server, if necessary.

# smtp_password:
# Password to log in to your smtp server. Must be specified for
# smtp_login to have an effect.

# smtp_port:
# Port number for smtp on smtp_host. Defaults to 0.

# smtp_ssl:
# If true, connects to smtp through SSL. Defaults to false.

# smtp_timeout:
# Optionally sets the number of seconds after which smtp attempts should
# time out.

worker-count-uniques: false
# If true, workers will only count unique pending jobs when deciding
# whether to stay alive. So if a worker can't get a job to run and other
# workers are waiting on all of its pending jobs, the worker will die.
# worker-keep-alive must be true for this to have any effect. Defaults
# to false.

worker-keep-alive: false
# If true, workers will stay alive when they run out of jobs to run, as
# long as they have some pending job waiting to be run. Defaults to false.

worker-ping-interval: 1.0
# Number of seconds to wait between pinging scheduler to let it know
# that the worker is still alive. Defaults to 1.0.

# worker-task-limit:
# .. versionadded:: 1.0.25
#
# Maximum number of tasks to schedule per invocation. Upon exceeding it,
# the worker will issue a warning and proceed with the workflow obtained
# thus far. Prevents incidents due to spamming of the scheduler, usually
# accidental. Default: no limit.

# worker-timeout:
# .. versionadded:: 1.0.20
#
# Number of seconds after which to kill a task which has been running
# for too long. This provides a default value for all tasks, which can
# be overridden by setting the worker-timeout property in any task. This
# only works when using multiple workers, as the timeout is implemented
# by killing worker subprocesses. Default value is 0, meaning no
# timeout.

# worker-wait-interval:
# Number of seconds for the worker to wait before asking the scheduler
# for another job after the scheduler has said that it does not have any
# available jobs.

[email]

# These parameters control sending error e-mails through Amazon SES.

AWS_ACCESS_KEY: {{ aws_access_key | default(None) }}
# Your AWS access key

AWS_SECRET_KEY: {{ aws_secret_key | default(None) }}
# Your AWS secret key

region: {{ aws_region | default('us-east-1') }}
# Your AWS region. Defaults to us-east-1.

type: {{ luigi_email_type | default(None) }}
# If set to "ses", error e-mails will be send through Amazon SES.
# Otherwise, e-mails are sent via smtp.

[scheduler]

# Parameters controlling scheduler behavior

# disable-hard-timeout:
# Hard time limit after which tasks will be disabled by the server if
# they fail again, in seconds. It will disable the task if it fails
# **again** after this amount of time. E.g. if this was set to 600
# (i.e. 10 minutes), and the task first failed at 10:00am, the task would
# be disabled if it failed again any time after 10:10am. Note: This setting
# does not consider the values of the `disable-num-failures` or
# `disable-window-seconds` settings.

# disable-num-failures:
# Number of times a task can fail within disable-window-seconds before
# the scheduler will automatically disable it. If not set, the scheduler
# will not automatically disable jobs.

# disable-persist-seconds:
# Number of seconds for which an automatic scheduler disable lasts.
# Defaults to 86400 (1 day).

# disable-window-seconds:
# Number of seconds during which disable-num-failures failures must
# occur in order for an automatic disable by the scheduler. The
# scheduler forgets about disables that have occurred longer ago than
# this amount of time. Defaults to 3600 (1 hour).

record_task_history: true
# If true, stores task history in a database. Defaults to false.

# remove-delay:
# Number of seconds to wait before removing a task that has no
# stakeholders. Defaults to 600 (10 minutes).

# retry-delay:
# Number of seconds to wait after a task failure to mark it pending
# again. Defaults to 900 (15 minutes).

state-path: {{ luigi_state_file }}

# Path in which to store the luigi scheduler's state. When the scheduler
# is shut down, its state is stored in this path. The scheduler must be
# shut down cleanly for this to work, usually with a kill command. If
# the kill command includes the -9 flag, the scheduler will not be able
# to save its state. When the scheduler is started, it will load the
# state from this path if it exists. This will restore all scheduled
# jobs and other state from when the scheduler last shut down.
# Sometimes this path must be deleted when restarting the scheduler
# after upgrading luigi, as old state files can become incompatible
# with the new scheduler. When this happens, all workers should be
# restarted after the scheduler both to become compatible with the
# updated code and to reschedule the jobs that the scheduler has now
# forgotten about.
# This defaults to /var/lib/luigi-server/state.pickle

# worker-disconnect-delay::
# Number of seconds to wait after a worker has stopped pinging the
# scheduler before removing it and marking all of its running tasks as
# failed. Defaults to 60.

[task_history]

# Parameters controlling storage of task history in a database

db_connection: {{ luigi_history_db }}
# Connection string for connecting to the task history db using
# sqlalchemy.

AWS_ACCESS_KEY: {{ aws_access_key | default(None) }}

AWS_SECRET_KEY: {{ aws_secret_key | default(None) }}

region: {{ aws_region | default('us-east-1') }}

type: {{ luigi_email_type | default(None) }}

where did you found those parameters? I can't find them in the luigi documentation