Measure endpoint latency
Documentation: Microsoft Learn
The loop sends four ICMP probes per target with Test-Connection and pipes the replies into Measure-Object, which computes the average, minimum, and maximum response time.
The script
Script reader
Inspect before you run
Inspect the script below. It does not run in this browser.
Inspection only — no command has run.
Show reference output
PowerShell 7 only. This example measures the Latency property returned by PowerShell 7's Test-Connection. On Windows PowerShell 5.1 the equivalent property is ResponseTime — on 5.1 this script fails instead: Test-Connection returns Win32_PingStatus objects, which have no Latency property, so Measure-Object Latency reports the error Property "Latency" cannot be found in any object(s) input. Swap Latency for ResponseTime to run it on 5.1.
Inspect this example’s structure
Read the command and its supplied arguments. This example requires real PowerShell or its module; this diagram does not execute it.
$Targets = "SRV01", "SRV02"
foreach($Target in $Targets){ Test-Connection $Target -Count 4 | Measure-Object Latency -Average -Minimum -Maximum | Select-Object @{ Name = "Target"; Expression = { $Target } },Average,Minimum,Maximum }- Stage 1$Targets
Positional values or no named parameters
How it works
The loop sends four ICMP probes per target with Test-Connection and pipes the replies into Measure-Object, which computes the average, minimum, and maximum response time. A calculated property stamps the target name onto each summary row so the results stay attributable.
Adapting it
Raise -Count for a longer sample when the problem comes and goes, and add -Delay to spread the probes out rather than firing them back to back into the same queue.
Requirements
ICMP permitted to the targets, which plenty of hardened environments block outright.
If the result is unexpected
A latency report shows failures or inconsistent times despite a working application.
ICMP may be blocked or deprioritized, and a small sample does not represent application response time.
- Separate ping loss and round-trip time from a test of the actual application endpoint.
Validate the result
The report states protocol, sample count and failures instead of averaging failures into a valid time.
Keep the check reliable
Use a repeatable measurement window and compare like-for-like network paths.
