ESXTOP Storage trableshooting

Storage Metrics in ESXTOP: Understanding Good and Bad Statistics


ESXTOP provides a wealth of storage metrics to help administrators gauge the health and performance of their storage subsystems. Let’s delve into some key storage metrics, distinguishing between good and bad statistics.

1. Device Latency (DAVG):

  • Good: Low and consistent values (e.g., <10 ms) indicate efficient storage operations.
  • Bad: Spikes or sustained high values suggest potential storage bottlenecks, impacting VM performance.

2. Kernel Latency (KAVG):

  • Good: Low and stable values (e.g., <2 ms) indicate that the ESXi kernel is efficiently processing storage commands.
  • Bad: Elevated values may indicate issues in the ESXi host or storage subsystem.

3. Queue Depth (QUED):

  • Good: A reasonable queue depth (e.g., <16) implies efficient storage handling.
  • Bad: High queue depth may lead to performance degradation, indicating that the storage system is struggling to manage the workload.

4. Commands per Second (CMDS/s):

  • Good: Steady and moderate values signify a healthy rate of storage commands.
  • Bad: Drastic fluctuations or consistently high values may indicate stress on the storage subsystem.

5. Read and Write Data Rates (MBREAD/s, MBWRTN/s):

  • Good: Balanced and steady rates for reads and writes suggest a well-distributed workload.
  • Bad: Significant disparities or erratic patterns may indicate workload imbalances or inefficient storage utilization.

6. Bus Resets (RESETS/s):

  • Good: Low or zero values indicate a stable storage bus.
  • Bad: Frequent bus resets may suggest issues with the storage connection, potentially impacting reliability.

7. Device Errors (DAES/s):

  • Good: Minimal or zero errors suggest a stable and reliable storage environment.
  • Bad: Increasing error rates may indicate hardware issues or a problematic storage subsystem.

8. Consolidation Ratio (CONS/s):

  • Good: Low consolidation ratios indicate efficient use of storage resources.
  • Bad: High consolidation ratios may lead to contention, impacting storage performance.

9. Commands Aborted (ABRTS/s):

  • Good: Low or zero aborted commands indicate a stable storage environment.
  • Bad: Increasing aborted commands may suggest issues with storage connectivity or misconfigurations.


Regularly monitoring these storage metrics in ESXTOP provides administrators with valuable insights into the health and performance of their storage infrastructure. By understanding what constitutes good or bad statistics, administrators can proactively address issues, optimize performance, and ensure the reliability of their virtualized environments.




**1. Launch ESXTOP:

  • Command: Open a terminal and run esxtop.
  • Navigation: Press u for disk view.

**2. Identify Storage Devices:

  • Command: Press d to display disk-related statistics.
  • Focus: Identify storage devices by their names (e.g., naa.XXXXXXXXXXXX).

**3. Check Device Latency:

  • Metric: Look at the “DAVG/cmd” column.
  • Threshold: Lower values are better; high values indicate latency issues.

**4. Monitor Queue Depth:

  • Metric: Observe “QUED” and “ACTV” columns.
  • Threshold: High queue depth or active values may indicate resource contention.

**5. Evaluate Throughput:

  • Metric: Focus on “KAVG” and “GAVG” columns.
  • Threshold: High values may indicate I/O congestion; compare with storage array capabilities.

**6. Review Read and Write Rates:

  • Metric: Examine “KAVG/rd” and “KAVG/wr” columns.
  • Threshold: High rates may indicate intensive read or write operations.

**7. Check Device Utilization:

  • Metric: Look at “%UTIL” column.
  • Threshold: High values suggest the device is saturated; check for bottlenecks.

**8. Identify Storage Paths:

  • Command: Press n for the NMP (Native Multipathing) view.
  • Focus: Recognize paths and check their status.

**9. Check Path Latency:

  • Metric: Look at “DCTRS” column.
  • Threshold: Elevated values indicate path latency; investigate further.

**10. Review Datastore Latency:

  • Command: Press v for datastore view.
  • Metric: Check “DAVG/cmd” and “KAVG/cmd.”
  • Threshold: High values may indicate datastore latency.

**11. Monitor Datastore Performance:

  • Metric: Observe “ABRT/s” (Aborts per second) and “DGRP/s” (Device grips per second).
  • Threshold: High abort rates suggest issues; investigate causes.

**12. Utilize VMware Knowledge Base:

  • Resource: VMware’s Knowledge Base provides solutions to common storage issues.

**13. Community Forums and Documentation:

  • Resource: VMware community forums and official documentation offer valuable insights.

**14. Backup Before Changes:

  • Important: Always back up critical data before implementing changes.

**15. Document Changes:

  • Best Practice: Keep a record of changes made during troubleshooting for future reference.


Empowering beginners to navigate ESXTOP and interpret storage metrics is crucial for maintaining optimal virtualized environments. Regularly monitoring these metrics, understanding thresholds, and taking proactive measures will contribute to efficient storage management and troubleshooting.