Understanding memory usage in your application
Understanding memory usage in your application
How memory statistics work
Your application's memory usage is monitored in real-time and displayed in the statistics tab. The system checks memory consumption every 5 seconds, giving you an up-to-date view of how your application is performing.
Reading the memory indicators
Memory usage is shown with three visual elements:
Percentage: Shows how much of your allocated memory is currently being used (e.g., "45.2%")
Progress bar: A visual representation of memory consumption with color coding
Actual usage: Displays the exact amount used versus allocated (e.g., "512Mi / 1Gi")
Understanding the color codes
The memory indicator uses traffic light colors to help you quickly assess usage:
Green (0-59%): Your application has plenty of memory available. This is the ideal operating range.
Yellow (60-79%): Memory usage is moderate. Your application is functioning well but you should monitor for increases.
Red (80%+): High memory usage. While your application may still run fine, you're approaching the limit.
Is high memory usage always a problem?
No, high memory usage isn't necessarily bad. Here's why:
Applications often use available memory for caching to improve performance
Memory usage naturally fluctuates based on user activity and workload
Brief spikes to 80-90% during peak times are normal
Your application has a 10% buffer above its guaranteed memory for handling temporary spikes
When should you consider upgrading?
Look for these signs that indicate you need more memory:
Consistent red indicators: If memory stays above 80% for extended periods
Application crashes: If your application restarts with exit code 137, it ran out of memory
Performance issues: Slow response times combined with high memory usage
Multiple services at high usage: When both your application and database show high memory consumption
How memory allocation works
When you select memory for your application:
Your application is guaranteed 90% of the selected amount at all times
The remaining 10% serves as a buffer for temporary spikes
Memory is automatically calculated based on CPU selection
Each instance gets the full memory allocation when running multiple instances
Taking action
If you determine you need more memory:
Go to the Resources tab in your application settings
Adjust the CPU slider (memory increases automatically with CPU)
Review the updated monthly cost
Save your changes to apply the new allocation
Pro tip: Monitor your application's memory patterns over a week to understand its true needs before upgrading. Short-term spikes during deployments or batch processes are normal and don't require permanent resource increases.