Hi everyone! I'm still pretty new to AWS and have been focusing on cost optimization, so I wanted to share my experience and get your thoughts.
I recently went through and tagged all my AWS resources (EC2 instances, S3 buckets, RDS, you name it) with detailed labels for `environment`, `project`, `owner`, and `cost-center`. My goal was to get better visibility into spending and, honestly, to make our dashboards in CloudWatch more useful.
The result was a mixed bag:
* **Insights improved:** My Cost Explorer reports are now amazing! I can finally see exactly which project is blowing the budget. Creating custom dashboards for each team is way easier.
* **But... my CloudWatch bill literally quintupled the next month.** 😳
I think I understand *why*—more tags means higher cardinality on my metrics, which increases my CloudWatch bill—but I wasn't prepared for a 5x jump. It feels like I traded one cost problem for another.
Has anyone else run into this? I'm looking for advice on:
* Is there a way to keep the detailed tagging for cost allocation without getting hammered on observability costs?
* Do other observability platforms handle high-cardinality data more cost-effectively than CloudWatch?
* Should I be more selective about which resources get the full tag treatment?
I love the clarity the tags provide, but not at this price. Any tips or shared experiences would be really helpful!
Every dollar counts.
Oof, that's a huge jump! I was thinking of doing the same for our email infra. Thanks for the heads up.
> I think I understand why - more tags means higher cardinality
Exactly. I've heard you can set up metric filters to only publish the high-cardinality stuff you *really* need to CloudWatch. Maybe log that super granular data elsewhere?
Which observability tools are you considering instead? I'm looking at a few but the pricing pages are always so vague.