Reliability
Latency Stability Test: Track Jitter Over Time
The Latency Stability Test runs sustained ping monitoring to characterize how steady your connection is over time. Where a basic jitter test captures momentary variance, this tool detects multi second drift, recurring spikes, and slow degradations that only appear over longer windows.
What It Measures
This tool measures latency stability by sampling ping at fixed intervals and computing variance, standard deviation, drift, and spike count over the full observation window. It also calculates a time series chart so you can see exactly when problems occurred.
How It Works
- Probes the nearest edge server at a fixed cadence (typically 1 to 2 samples per second)
- Records each sample with a timestamp for time series analysis
- Calculates rolling jitter, drift, and spike count windows
- Outputs a stability grade and a chart showing your full latency history
Why It Matters
A stable connection feels predictable. An unstable one feels broken even when speed tests look fine. Voice calls, gaming, and remote desktop all suffer when latency wanders or spikes. Long form stability monitoring catches problems that 5 second tests cannot, like WiFi roaming, ISP queue cycles, and thermal throttling on cable modems.
Understanding Your Results
Standard deviation under 2ms over a 60 second window is excellent. Under 5ms is good. Above 10ms indicates noticeable instability. Zero spikes (samples more than 3x median) is ideal. Drift under 1ms over the test window means your baseline is steady.
Frequently Asked Questions
What counts as a latency spike?
A latency spike is any sample significantly above the rolling median, typically 3 to 5 times higher. A connection with a 20ms median that occasionally hits 100ms has spikes. Even infrequent spikes degrade real time applications because they happen exactly when the network is under load, which is when you notice.
What is latency drift?
Latency drift is a slow change in your baseline ping over time, separate from short jitter. If your median ping starts at 18ms and slowly climbs to 35ms over 5 minutes, that is drift. Drift often indicates upstream congestion building, a router queue filling up, or thermal effects on cable modems.
How long should I run the test?
Run for at least 60 seconds for basic stability profiling. Run for 5 to 10 minutes if you want to catch periodic issues like WiFi channel hopping, ISP traffic shaping cycles, or background application activity. Longer runs produce more reliable variance and drift measurements.
What if my latency is stable but high?
High stable latency means your routing is consistent but the path is long, often because your ISP routes you through a distant peering point. The fix involves contacting your ISP about routing or choosing a closer server when possible. Stable high latency is still better than unstable low latency for most applications.
Related Tools
Browse all tools
Explore all 50+ network tools, from speed and latency to DNS, routing, and CDN diagnostics. Free, instant, no installation required.
See all network tools