Fixing Stale or Incorrect Metric Values

When Server Scout displays stale or incorrect metric values, there are several common causes and straightforward fixes. This guide walks you through diagnosing and resolving the most frequent issues.

Check if the Agent is Running

The most common cause of stale metrics is a stopped agent. First, verify the agent service is active:

sudo systemctl status scout-agent

If the service is inactive or failed, restart it:

sudo systemctl restart scout-agent

If you see errors in the status output, check the agent logs for more details:

sudo journalctl -u scout-agent -f

Also verify in the Server Scout dashboard that your server isn't paused. Paused servers won't collect or display updated metrics.

Optional Metrics Showing Null Values

Some metrics may appear as null or missing entirely. This typically happens for two reasons:

1. Metric not enabled in configuration

Optional metrics need to be explicitly enabled. Run the configuration utility:

sudo /opt/scout-agent/scout-agent.sh --configure

Review the available options and enable any metrics you need.

2. Required command not installed

Some metrics depend on system tools that may not be installed. For example, disk I/O statistics require iostat. Install missing packages:

# On Ubuntu/Debian
sudo apt install sysstat

# On CentOS/RHEL
sudo yum install sysstat

Understanding Slow-Updating Metrics

Server Scout uses a 5-tier data collection system, and some metrics update less frequently than others:

• Service status: Updates hourly (glacial tier)
• Package updates: Checked daily
• Plugin data: Refreshes hourly
• Basic system metrics: Update every 5-30 seconds

This is by design to balance monitoring coverage with system resource usage. If you need immediate updates, force a collection cycle:

sudo /opt/scout-agent/scout-agent.sh --refresh

Memory Values Appearing Incorrect

Linux memory reporting often confuses users because the system uses available RAM for disk caching, which improves performance. This means "used" memory includes cache memory.

Server Scout displays both metrics:

• Used memory: Includes system cache (may appear high)
• Available memory: Shows truly available RAM (the important metric)

Focus on the "available" memory metric for genuine memory pressure. High "used" memory with adequate "available" memory is normal and healthy Linux behaviour.

Clock Synchronisation Issues

If timestamps appear incorrect or metrics seem delayed, check your server's clock synchronisation:

# Check current time sync status
timedatectl status

# Enable NTP synchronisation if disabled
sudo timedatectl set-ntp true

Servers with clocks more than a few minutes out of sync may display confusing timestamps. Ensure NTP is working properly:

# Check NTP sync status
chrony sources -v
# or on older systems:
ntpq -p

Data from Spool Replay

After connectivity outages, the agent replays spooled data to fill historical gaps in the dashboard. This may cause older timestamps to appear with new data points.

This behaviour is normal and expected—it ensures you have complete historical data even after temporary network issues. The agent automatically clears spooled data after successful transmission.

Force a Complete Refresh

To immediately update all metrics across all collection tiers:

sudo /opt/scout-agent/scout-agent.sh --refresh

This forces the agent to run fast, medium, slow, glacial, and daily collection cycles immediately. New data should appear in the dashboard within a few minutes.

Additional Diagnostics

If problems persist after these steps:

1. Check the agent's configuration file: /opt/scout-agent/scout.conf
2. Review recent system changes that might affect monitoring
3. Verify network connectivity to Server Scout's servers
4. Check system resource availability (disk space, memory)

For complex issues, our AI support bot can provide targeted troubleshooting within approximately 1 minute. Submit a support ticket through the dashboard with specific details about which metrics are problematic and any error messages you've encountered.

The combination of proper agent status, correct configuration, and understanding Linux's memory management typically resolves most metric display issues quickly.