tokencounter

README

Service tokencounter

Short summary

The tokencounter service is a processor (micro-service) that runs by default in the Gecholog container. tokencounter extracts token consumption fields from the response payload and creates a new field in the log that enables standardized reporting across different models. tokencounter can also throttle the traffic based on token consumption.

Options

Option	Description
-a	alias, set the name of the service
-o	specify filepath to configuration file
--validate	print config validation info (accepts stdin)
--version	print version

Example

This is how to validate a new config file in the docker context

docker exec -e NATS_TOKEN=set gecholog ./tokencounter -o app/conf/new_tokencounter_config.json --validate

Configuration file

Example of the configuration file for the tokencounter service can be found here.

Field	Description
cap_period_seconds	period in seconds after which token cap counter resets
log_level	one of DEBUG INFO WARN ERROR
service_bus	internal service bus configuration
token_caps	array token caps
usage_fields	array of patterns for tokens
version	the config file conforms to this specification

Usage field settings

Field	Description
router	specify router where pattern applies, or default
patterns	arrays of search patterns

Field	Description
field	field name
pattern	json-search path to token object

json-search path uses gjson syntax.

Token cap settings

Field	Description
router	specify router where cap applies
fields	arrays of cap fields

Field	Description
field	field name
value	token cap. 0 means no cap