Adding parameters to a pyspark job in AWS Glue

gistfile1.txt

from awsglue.transforms import *
from awsglue.utils import getResolvedOptions
from pyspark.context import SparkContext
from awsglue.context import GlueContext
from awsglue.job import Job
from pyspark.sql.types import *
from awsglue.dynamicframe import DynamicFrame

## @params: [JOB_NAME, CUSTOM1, CUSTOM2, CUSTOM3]
args = getResolvedOptions(sys.argv, ['JOB_NAME', 'CUSTOM1', 'CUSTOM2', 'CUSTOM3'])

# use args as usual:
print "Custom 1 is {}".format(args['CUSTOM1'])

NOTE that in order for this to work, the key-value pairs you trigger your job with should be of the form:
--CUSTOM1 value1
--CUSTOM2 value2

Note the double-dash option specifier; this is required by the 'getResolvedOptions' function.
This was non-obvious from the docs.