Skip to main content

Remediating ElastiCache Clusters with Low Hit Rates

Overview​

This tutorial demonstrates how to identify and remediate ElastiCache clusters with low cache hit rates. A cache hit rate below 80% typically indicates the cache isn't providing sufficient value relative to its cost. You'll learn how to evaluate cache metrics and safely delete underperforming clusters.

What you'll accomplish:

  • Navigate to the ElastiCache console
  • Review cache hit rate metrics
  • Delete a low-performing cache cluster

Prerequisites:

  • AWS Console access with ElastiCache permissions
  • An ElastiCache cluster to evaluate

Step 1: Navigate to ElastiCache Dashboard​

Navigate to the ElastiCache console in us-east-1. The Redis clusters view shows all active cache clusters in your account.

ElastiCache Dashboard

Step 2: Access Cluster Details​

Click on the cluster remediation-elasticache-low-hit-rate-redis to view its details.

Step 3: View Cache Metrics​

Click the Metrics tab to view cache performance. Look for the CacheHitRate metric over the past 7 days.

A healthy cache should maintain >80% hit rate. If the hit rate is consistently below this threshold, the cache may not be cost-effective.

Cluster Metrics

Step 4: Evaluate Hit Rate Data​

Note: New clusters won't show meaningful hit rate metrics until they've handled production traffic for several days. For this tutorial cluster, metrics will populate after 30 days of operation.

Decision criteria:

  • Hit rate <50%: Strong candidate for elimination
  • Hit rate 50-80%: Investigate optimization opportunities
  • Hit rate >80%: Cache is performing well

Step 5: Return to Cluster List​

Return to the Redis clusters list to delete the underperforming cache.

Step 6: Initiate Cluster Deletion​

Select the cluster remediation-elasticache-low-hit-rate-redis and click Delete from the Actions menu.

Delete Cluster

Step 7: Configure Deletion Options​

Deletion options:

  • Create final backup: Choose "No" for clusters being eliminated due to low value. Choose "Yes" if there's any possibility you'll recreate the cache with the same data.
  • Confirmation: Type the cluster name to confirm deletion.

Click Delete to proceed.

Deletion Confirmation

Step 8: Verify Deletion​

The cluster will show a status of deleting. Deletion typically completes within 5-10 minutes. The cluster will disappear from the list once deletion is complete.

Cost Impact​

Deleting a cache.t3.micro cluster saves approximately $12/month. Larger instances (cache.m5.large, cache.r5.xlarge) can save $100-500+/month. Savings are realized immediately upon deletion.

Alternative Approaches​

Before deleting, consider these optimization strategies:

  • Optimize cache strategy: Review cache key patterns, increase TTL for stable data, implement cache warming, or resize to match working set
  • Application-level caching: Consider using in-memory caches in your application servers instead of ElastiCache
  • Monitor first: If uncertain, monitor for 1-2 weeks after optimizations before deciding to delete
  • Infrastructure as Code: For IaC-managed resources, delete through your infrastructure code (CloudFormation/Terraform) rather than the console

Summary​

You've successfully identified and deleted an ElastiCache cluster with low cache hit rates. This remediation eliminates ongoing costs for underperforming infrastructure and helps maintain a cost-effective AWS environment.