Managing local disk write latency in Confluence Data Center

This insight appears when write operations to your local disk write latency directory exceed 5ms. Local write latency measures how quickly each node can write data to its local home directory. This metric is crucial for operations like search indexing, plugin functionality, and temporary file management.

How does high local disk latency affect performance?

When local disk operations slow down, users and administrators experience several significant impacts:

• Search Functionality

  • Delayed index updates mean search results may be outdated

  • Slower search response times across all content types

  • Extended duration for full reindex operations

  • Inconsistent quick search results

• Plugin Operations

  • Increased plugin installation times

  • Slower plugin startup during instance restarts

  • Delayed plugin data storage operations

  • Potential plugin timeout issues

• General System Operations

  • Slower temporary file handling

  • Delayed log file writes

  • Reduced overall system responsiveness

  • Potential cascade effect on memory usage

What's the recommendation?

To improve local disk write latency, focus on these key areas:

Monitor I/O Performance Patterns

Start by establishing a clear picture of your disk activity:

• On Linux Systems:

  
  

  # Monitor disk I/O statistics 
  iostat -x 1 
  # Track top I/O processes 
  iotop -o

  
  

  
  

• On Windows Systems:

  • Use Performance Monitor (perfmon.exe)

  • Focus on Disk Queue Length

  • Monitor Avg. Disk sec/Transfer

  • Track Active Disk Time

Look for patterns in these metrics that correlate with performance issues. Pay particular attention to:

• Consistent queue lengths above 2

• Disk utilization consistently above 80%

• Spikes in write latency during specific operations

Learn more about performance tuning

Verify disk health

Ensure storage system health:

Verify disk health and configuration

Regular storage system maintenance is crucial for optimal performance:

• Storage Health

  • Monitor RAID health, review error logs, and track performance trends regularly.

  • Verify disk status and filesystem health as part of your routine checks.

• Performance Configuration Ensure your storage setup follows best practices:

  • Use recommended storage types for your deployment

  • Implement appropriate RAID levels

  • Configure optimal filesystem and mount options

Review system requirements

Optimize system configuration

Address resource constraints:

Process management

Coordinate competing processes that might impact disk performance:

• Schedule antivirus scanning during off-peak hours

• Plan backup operations to minimize impact

• Coordinate system maintenance windows

• Resource Allocation :

  • Balance workloads across your cluster and maintain appropriate resource allocation for critical processes.

Learn about Data Center requirements

Preventive Measures

Implement these ongoing practices to maintain optimal disk performance:

• Set up automated monitoring for latency thresholds

• Schedule regular maintenance windows for disk optimization

• Document baseline performance metrics

• Review and update storage firmware regularly

Monitor application performance