VMware monitoring must haves

VMware performance monitoring -

The Must Haves

OpManager Marketing Team

Agenda

• About ManageEngine• Virtualized vs. Physical Server Monitoring• VMware Performance Monitors• Performance Management – the Essentials• OpManager’s VMware Monitoring• Discussion

About ManageEngine

ManageEngine Software is used by over 40,000 customers including 3 out of every 5 Fortune 500 companiesManageEngine Software is used by over 40,000 customers including 3 out of every 5 Fortune 500 companies

Vendor Landscape

ManageEngine – IT Management Portfolio

ManageEngine is the only IT management vendor focused on bringing a complete IT management portfolio to the mid sized enterpriseManageEngine is the only IT management vendor focused on bringing a complete IT management portfolio to the mid sized enterprise

Pros Cons

Shared Resources Ideal resource utilization, quickest ROI

Problem in one VM might affect the other VMs in the Host

Dynamic Servers 100% server uptime Tracking of VMs to corresponding Host

Easy Provisioning Server provisioning within minutes

VM Sprawl – Zombie VMs eating-up resources

Virtualized vs. Physical Server Monitoring

Performance Monitoring – the Fundamentals

VMware Performance Monitors

For Hosts & VMs: CPU Utilization (%), CPU Usage(MHz) On Individual VMs: CPU Used (ms), CPU Ready time (ms), CPU Wait time (ms)

CPU

• CPU Ready time : VM ready-to-run but no physical CPU free• CPU Wait time : VM blocked on I/O• Rule of thumb: CPU ready time >20%, needs investigation• High %ready + high %used, very probable CPU over-commitment

EMA benchmark report – Avg. performing enterprise has physical EMA benchmark report – Avg. performing enterprise has physical CPU utilization of 45%, best performing enterprise has it at 70%CPU utilization of 45%, best performing enterprise has it at 70%

Memory

For Hosts & VMs - Memory Utilization (%)/ Usage (KB), Memory Active (KB), Used (KB) & Overhead (KB)On Hosts - Swap In/ Out (KB), Swap used (KB), Swap in/ Out rate (KB/s), memory shared (KB), shared common(KB)On individual VMs - Memory Balloon (KB), Shared(KB) & Swapped (KB)

• Rule of thumb: >1MB/s swap in or swap out rate, memory over-commitment, reallocation needed• Memory shared – shared common = machine memory savings (KB)

EMA report benchmark – Avg. performing enterprise has EMA report benchmark – Avg. performing enterprise has memory utilization of 60%, best performing enterprise 80%memory utilization of 60%, best performing enterprise 80%

VMware Performance Monitors

Hard Disk

For Hosts & VMs - Disk I/O Usage (KBps), Disk Read & Write Speed (KBps)/ Disk Read & Write Speed Requests (number), Disk Bus ResetOn Hosts - Disk Read/ Write Latency (ms), Disk Command Abort, Device, Kernel & Command Latency (ms)

• When Command Latency (total latency for one command) > 50 ms, high latency, needs investigation• If Kernel Latency (Average latency in vmkernel) very low ~0, command is not queuing in the kernel. If also Device Latency (Average latency at device) is 0, disk has connectivity issues.• Study Reads and Writes, CPU usage side-by-side

VMware Performance Monitors

Network Interface Card

For Hosts & VMs - Packets Received/ Transmitted (packets/sec), Network Usage (KBps), Network Received/ Transmitted Speed (KBps)

• Are packet rate/ bandwidth at expected levels?• Are the VMs sharing one physical NIC?

EMA benchmark report – Avg. performing enterprise has NIC EMA benchmark report – Avg. performing enterprise has NIC utilization of 30-40%, best performing enterprise have it at 70-utilization of 30-40%, best performing enterprise have it at 70-90%90%

VMware Performance Monitors

Performance Management – the Essentials

• Quick views - of problem areas e.g. top VMs by CPU Ready time• Thresholds - assigned by default for important metrics• Notifications & Alarm Management - should easily gel with the existing IT mgmt procedures• Side-by-side reports – a single dashboard having multiple metrics; VMware specific and those of the physical hardware for quick troubleshooting• History Reporting

OpManager’s VMware Monitoring

• Supports ESX 3.5, ESX3.5i, ESX4 & ESX4i • Uses native VMware APIs: Agent-less, in-depth monitoring, zero configurations• Over 70 deep metrics, with daily/weekly/monthly history reports; 15+ monitors with default thresholds • Automatically maps VMotioned VMs to the corresponding Hosts• Inbuilt with OpManager – no extra plug-in or download required• Exclusive dashboards & monitors for MS SQL, MS Exchange & Active Directory

OpManager‘s VMware Monitoring

VMware Monitoring Dashboard

OpManager‘s VMware Monitoring

VMware ESX Snapshot page

ManageEngine OpManager

www.opmanager.com

[email protected]

Discussion

Thanks

[email protected]