Prometheus exporter for AWS CloudWatch - Discovers services through AWS tags, gets CloudWatch metrics data and provides them as Prometheus metrics with AWS tags as labels
Current Behavior
Assume we run yace with "--scraping-interval=60" parameter with the specified config above. In such configuration GMD pricing gives about 6 dollars overhead daily comparing with the basic java cw exporter with the same config and GetMetricStatistics, which have 1kk free-tier API calls monthly.
On some comprehensive configs we can observe about 30 dollars daily overhead.
Expected Behavior
Steps To Reproduce
Run yace exporter with any config, than run java cw exporter with the same resources/dimension/metrics specified and compare pricing.
Anything else?
We wonder if such pricing overhead is expected behavior? Why the pricing information/comparisions are not specified in exporter readme? What else can be done to cost cut, except of reducing scraping interval and narrowing down list of scraping resources?
YACE version v0.59.0
Config file ` jobs:
type: AWS/RDS regions: [ eu-west-1 ] delay: 0 period: 60 length: 300 dimensionNameRequirements:
type: AWS/ApplicationELB regions:
type: AWS/ApplicationELB regions:
type: AWS/CloudFront regions: [ us-east-1 ] delay: 600 period: 60 length: 60 nilToZero: true dimensionNameRequirements:
type: AWS/Lambda regions:
type: AWS/Firehose regions: [ eu-west-1 ] delay: 0 period: 60 length: 300 nilToZero: true dimensionNameRequirements:
type: AWS/SQS regions: [ eu-west-1 ] delay: 0 period: 60 length: 300 nilToZero: true dimensionNameRequirements:
type: AWS/S3 regions:
Current Behavior Assume we run yace with "--scraping-interval=60" parameter with the specified config above. In such configuration GMD pricing gives about 6 dollars overhead daily comparing with the basic java cw exporter with the same config and GetMetricStatistics, which have 1kk free-tier API calls monthly. On some comprehensive configs we can observe about 30 dollars daily overhead.
Expected Behavior
Steps To Reproduce Run yace exporter with any config, than run java cw exporter with the same resources/dimension/metrics specified and compare pricing.
Anything else? We wonder if such pricing overhead is expected behavior? Why the pricing information/comparisions are not specified in exporter readme? What else can be done to cost cut, except of reducing scraping interval and narrowing down list of scraping resources?