Throw away data points with known invalid values
to avoid the whole batch being rejected by CloudWatch API.
Signed-off-by: Sergei Ivanov <sergei_ivanov@mail.ru>
* Added optional metric filtering
* Configuration options to provide whitelist and blacklist
patterns for metrics, so that only a subset of available metrics
is exported from Prometheus to CloudWatch
Signed-off-by: Sergei Ivanov <sergei_ivanov@mail.ru>
* Regenerated README.md
Signed-off-by: Sergei Ivanov <sergei_ivanov@mail.ru>
* Adding ability to replace dimensions. This is useful to aggregate metrics across dimensions in the kube_state metrics namespace. The aggregation across dimensions is neccessary for adding CloudWatch alarms to kube-state metrics for the variable number of pods/container/nodes that will exist in our EKS clusters
* Don't add replacement if not configured
* Addressing PR comments: Add new flag and contributor
