MAX_DIMENSIONS should be 9, not 10

[dlenski]

Per the upstream docs:

A DimensionSet MUST NOT contain more than 9 dimension keys.

Currently, you're setting MAX_DIMENSIONS to 10:

aws-embedded-metrics-python/aws_embedded_metrics/constants.py

Line 15 in 658b2eb

MAX_DIMENSIONS = 10

… and then using that to truncate the dimension set to 10 entries when serializing, e.g.:

aws-embedded-metrics-python/aws_embedded_metrics/serializers/log_serializer.py

Lines 23 to 30 in 658b2eb

	def serialize(context: MetricsContext) -> str:
	dimension_keys = []
	dimensions_properties: Dict[str, str] = {}
	for dimension_set in context.get_dimensions():
	keys = list(dimension_set.keys())
	dimension_keys.append(keys[0:MAX_DIMENSIONS])
	dimensions_properties = {**dimensions_properties, **dimension_set}

I'm not sure that “silently truncating the dimension set” to an acceptable size is the best solution, but in any case the MAX_DIMENSIONS should be reduced to 9.

[jaredcnance]

Thanks for reporting this! Definitely open to proposals on how to treat this. I've considered adding a configuration parameter that would throw exceptions or to log and silently move on in these kinds of cases.

[dlenski]

Based on the Zen of Python (“Explicit is better than implicit”, “Errors should never pass silently unless explicitly silenced”, and all that) I would say that throwing an exception is probably the right default behavior.