Understanding Server Metrics History

Server Scout automatically collects and stores performance metrics from your monitored servers, maintaining a comprehensive historical record that proves invaluable for troubleshooting, trend analysis, and capacity planning. Understanding how to interpret and utilise this historical data will help you make informed decisions about your infrastructure.

How Server Scout Stores Historical Data

Server Scout employs a sophisticated data aggregation system that balances storage efficiency with data granularity. The system stores metrics at multiple time intervals:

• 1-minute intervals: Raw data points collected every minute for recent activity
• 5-minute intervals: Aggregated data for short-term trend analysis
• Hourly intervals: Consolidated metrics for daily and weekly overviews
• Daily intervals: Long-term data points ideal for capacity planning and monthly reporting

This hierarchical storage approach ensures you have detailed information for recent events whilst maintaining long-term trends without consuming excessive storage space. Older, more granular data is automatically rolled up into larger time buckets as it ages.

Accessing Historical Graphs

To view historical metrics for any server, navigate to the server detail page by clicking on the server name from your dashboard. The historical graphs are displayed prominently, showing key performance indicators including:

• CPU usage percentage
• Memory utilisation
• Disk I/O activity
• Network traffic
• System load averages

Each graph displays colour-coded lines representing different metrics, making it easy to correlate events across multiple performance indicators.

Selecting Time Ranges

Server Scout provides five pre-defined time ranges to help you focus on relevant data periods:

1. 1 hour (1h): Shows 1-minute granularity data, perfect for investigating recent performance spikes or current issues
2. 6 hours (6h): Displays 5-minute aggregated data, useful for understanding recent trends and patterns
3. 24 hours (24h): Presents hourly data points, ideal for identifying daily usage patterns
4. 7 days (7d): Shows daily aggregations, excellent for weekly trend analysis and comparing weekday vs. weekend performance
5. 30 days (30d): Displays daily summaries, essential for monthly capacity planning and long-term trend identification

To change the time range, simply click the desired period button above the graphs. The display will automatically refresh to show the selected timeframe.

Understanding Min/Max/Average Values

When viewing aggregated data (anything beyond 1-minute intervals), Server Scout displays three key statistical values:

Average: The mean value across all data points in the aggregation period. This provides a baseline understanding of typical performance during that timeframe.

Minimum: The lowest recorded value, which can help identify periods of low activity or underutilisation.

Maximum: The highest recorded value, crucial for identifying performance peaks, bottlenecks, or unusual spikes in activity.

These values appear as separate lines on the graphs when viewing longer time periods. For instance, when examining a 7-day period, each daily data point shows the minimum, maximum, and average values for that day.

Leveraging Historical Data for Pattern Recognition

Historical metrics excel at revealing patterns that aren't obvious during day-to-day monitoring:

• Daily patterns: Identify peak usage hours and plan maintenance during low-activity periods
• Weekly cycles: Recognise differences between weekday and weekend loads
• Seasonal trends: Spot longer-term changes in resource requirements
• Performance degradation: Detect gradual increases in response times or resource usage that might indicate underlying issues

Capacity Planning with Historical Data

Use the 30-day view to make informed infrastructure decisions:

# Example: If CPU averages show consistent 70%+ usage over 30 days,
# consider scaling up or optimising workloads

Monitor memory trends to predict when additional RAM might be needed, and track disk I/O patterns to identify storage bottlenecks before they impact performance.

Regular review of historical data, particularly during monthly infrastructure reviews, enables proactive rather than reactive server management. This approach helps maintain optimal performance whilst avoiding unnecessary over-provisioning of resources.