What Config Drift Costs You: CloudWatch Log Retention, S3 Lifecycle Policies and RDS Multi-AZ, Priced
Config drift usually gets discussed as a correctness problem: the running state no longer matches the intended one. But three specific drifts are money problems first, and they share a shape: a default nobody chose, set at provision time, growing silently ever since. None of them shows up as an incident. All of them show up on the bill, unlabeled. Here they are, priced, with the fix for each.…
Three specific drifts due to Config drift have monetary implications, all stemming from default settings established during provision time and growing silently over time. The first drift concerns CloudWatch log groups that fail to expire, retaining logs indefinitely at $0.03 per GB-month, regardless of the actual usage. This can result in substantial storage costs, such as a service logging 5 GB daily accumulating to approximately 3.6 TB in two years, costing around $110 monthly.
This issue can be addressed by setting a retention period of 30 to 90 days for most operational logs, after which the data automatically deletes itself.
The second drift focuses on S3 buckets lacking lifecycle policies, which keep data in the higher-cost Standard storage tier indefinitely. Standard storage costs $0.023 per GB-month, while Intelligent-Tiering can automatically transition objects to lower-cost tiers based on access patterns, saving significant money. For example, 10 TB of aging logs in Standard costs $230 per month, while the same data in Deep Archive would only cost about $10.
Furthermore, the lifecycle rules have specific duration requirements for each storage tier, ensuring that data is stored appropriately based on its usage frequency.
The third drift involves incorrectly applying Multi-AZ settings for RDS instances. While Multi-AZ doubles the instance and storage costs, it can also introduce risk if it is misapplied. Production databases should only have Multi-AZ if necessary, as a single-AZ setup can save costs but may be susceptible to outages due to AZ events.
On the other hand, non-production environments that have Multi-AZ enabled are experiencing cost drift, typically due to copied Terraform modules from production to non-production environments. These settings should be deliberately chosen per environment and periodically audited to ensure accuracy.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
