High Memory Usage Warning in Redis Cloud: What It Means and What to Do Next

Redis Cloud generates a high memory usage warning when your database approaches its configured memory limit. This usually means your dataset, indexes, or workload growth is pushing the database toward eviction or out-of-memory conditions. 

This article explains what the warning means, the risk level at different thresholds, and how to decide between Optimize, Increase Size, or Upgrade Plan. For deeper operational diagnostics, see Troubleshooting High Memory Usage in Redis Cloud.

What Triggers the Warning

You may see alerts such as:

• “High memory usage”

• “Dataset size exceeds X percent of plan limit”

• OOM errors in logs

Internally, this indicates:

• used_memory is approaching maxmemory

• Evictions may begin

• Writes may fail if eviction is not possible

By default, Redis Cloud also triggers a system alert email when memory usage reaches 80% of the database limit. This threshold is configurable in the Redis Cloud Console under Database Configuration → Alerts, where you can adjust or disable the alert based on your operational needs.

Risk Levels

• 80 percent
  Early warning. The growth trend should be reviewed.

• 85 to 95 percent
  Capacity pressure. Evictions or latency increases likely.

• Above 95 percent or OOM errors
  Capacity incident. Immediate action is required unless the eviction policy is expected and appropriate for your workload.

What Happens If You Ignore It

Impact depends on workload type.

Cache workloads

• Increased evictions

• Lower hit rate

• Higher latency

Persistent workloads

• OOM errors

• Write failures

If the eviction policy is set to noeviction, Redis will return OOM errors instead of evicting keys when memory limits are reached.

Application impact

High memory usage is not cosmetic. It indicates a real capacity or data lifecycle issue that can directly affect application stability and performance.

Step 1: Identify Your Workload Type

Before acting, determine:

Is this database primarily a cache?
Data can be safely evicted and rebuilt.

Or is it persistent / source of truth?
Data loss is not acceptable.

This distinction determines whether eviction is acceptable and how urgently you must scale.

Step 2: Choose Your Path

Option A: Optimize

Choose this if:

• Memory growth is unexpected

• Large unused keys exist

• TTLs are missing

• Indexes may be oversized

Typical actions:

• Add or reduce TTLs

• Remove unused namespaces

• Trim large collections

• Optimize Search or JSON indexes

For detailed steps, see:

• Troubleshooting High Memory Usage in Redis Cloud

• Undersized Redis Cloud Databases: How to Diagnose and Fix 

Option B: Increase Database Size

Choose this if:

• Data is required and cannot be reduced

• Growth is legitimate and sustained

• Persistent workload is approaching OOM

Increase size in the Redis Cloud Console:

Database → Edit → Adjust memory size

Plan limits apply depending on Essentials vs Pro.

Option C: Upgrade Plan (Essentials to Pro)

Choose this if:

• You are on a large Essentials database nearing its scaling ceiling

• You need clustering for horizontal scaling

• You need higher throughput limits

• You anticipate continued growth

Special Considerations

Auto Tiering

Even with Redis on Flash:

• All keys and metadata live in RAM

• Large values must load into RAM when accessed

Auto Tiering reduces pressure but does not remove RAM limits.

When to Act Immediately

Act now if:

• You see OOM errors

• Evictions are spiking unexpectedly

• Latency increases correlate with memory saturation

• Memory exceeds 90 percent and continues to rise

For advanced diagnostics, continue to: Troubleshooting High Memory Usage in Redis Cloud 

Summary

A high memory usage warning means your Redis Cloud database is approaching its configured memory limit. Depending on workload and growth pattern, the correct response is to Optimize, Increase Size, or Upgrade Plan. Acting early prevents OOM errors, eviction storms, and user-visible latency.