search

signal-ai / terraform-aws-cloudwatch-prometheus

A terraform module to provision the infrastructure to connect a cloudwatch metrics stream to prometheus remote write endpoints
Apache License 2.0
2 stars 0 forks source link

[QUESTION] How does this save cost? #1

Open Jks08 opened 1 year ago

Jks08 commented 1 year ago

I want to understand how your solution is more cost effective compared to exporters like YACE and CloudWatch Exporter because if I have over a million metrics per day, the cost of Kinesis data firehose and that of Lambda will be a lot. I ask this because I am currently having issues managing my cost when I use YACE and CloudWatch Exporter.

Any and all help is appreciated. Thank you.

evilr00t commented 3 months ago

Hi @Jks08,

Sorry for no reply for a long time - didn't see it.

We've made another revision of this lambda service. Now you can specify what metrics you're interested in.

i.e: 

included_filters = [
    # specific metrics
    {
      namespace    = "AWS/SQS",
      metric_names = ["ApproximateNumberOfMessagesVisible", "AgeOfOldestMessage", "NumberOfMessagesDeleted"],
    }
    # all metrics
    {
      namespace    = "AWS/Kinesis",
      metric_names = []
    }

With this approach we've noticed a drop in the cost with previous revision of around 30-40%.