Follow this guide to diagnose and mitigate performance-related alerts for Block Storage volumes. It covers immediate checks, deep diagnostics using Akamai Cloud Pulse, and when to escalate by opening a support ticket.

1. Understand the alert

When an alert fires, confirm the following details first:

• Which volumes triggered the alert.
• Metric type: latency, IOPS, or throughput.
• Severity of the alert.
• Application impact, such as timeouts, slow requests, or failures.
• Workload characteristics. Block storage performance varies significantly based on workload characteristics. Metrics such as I/O size, queue depth, and the number of attached disks on the Linode influence the effective performance of a volume.
  • I/O size. The average size of read and write operations directly affects performance. Small random I/Os (4–16 KiB) are typically IOPS-bound, while large sequential I/Os (256 KiB–1 MiB) are throughput-bound.
  • Queue depth. This represents the number of outstanding I/O operations for the top active devices. Achieving maximum disk performance generally requires a higher queue depth.
  • I/O latency. This measures the average read and write latency across all attached disks. It includes OS and file system overhead and is influenced by both I/O size and queue depth. Larger queue depths or small random I/Os naturally increase latency.

2. Initial diagnostics (quick checks)

Validate the timeframe

Match the alert window with your monitoring tools and application logs. Look for correlations with:

• Deployments
• Batch jobs
• Backup and restore activity
• Traffic spikes
• Related timeouts, failures, errors, and exceptions

Verify Linode and volume limits

Ensure that the workload is not reaching the configured limits of the Block Storage volume. Check:

• IOPS limits
• Throughput limits
• Number of volumes attached to a Linode

For default limits, see Block Storage performance limits. If the workload regularly peaks near limits, scaling up the volume is recommended.

Check guest-level resource usage

Inside the Linode, run system performance utilities such as iostat, vmstat, and blktrace to inspect the resource usage.

3. Analyze volume metrics with Akamai Cloud Pulse

Akamai Cloud Pulse provides fine-grained time-series metrics for each volume. Access your volumes’ metrics in Cloud Manager and identify any trends, compare current behavior with past patterns, and quickly detect when something isn't operating as expected.

1. Log in to Cloud Manager, expand Monitor in the main menu, and select Metrics.
2. Select Volumes > Storage Performance from the Dashboard dropdown list.
3. Select the Region and the affected Volumes.
4. Inspect the following in the dashboard:
   • Latency time series (read and write p50, p90, p95, and p99)
   • Volume operations charts
   • Volume transferred bytes charts

👍

See Block Storage metrics for a list of available metrics with their API equivalents, descriptions, and units. For more information on using Akamai Cloud Pulse dashboards, its filters and widgets, see Metrics dashboards.

4. Escalation: Create a support ticket

Open a support ticket if:

• The alert persists after troubleshooting.
• Akamai Cloud Pulse shows abnormal metrics.
• You suspect a region-level issue that is not specific to your workloads or Linode.