Packet Loss
A little packet loss can ruin a phone call while barely touching a file download. These guides explain where loss actually comes from and how much is acceptable for each kind of traffic. They cover how to measure it without fooling yourself and how to fix the causes you find.
Guides in this topic
- Where Packets Actually Die: A Layer-by-Layer TourCabling, buffers, duplex, Wi-Fi, ISP congestion, firewall drops — and the evidence each leaves behind.
- How Much Packet Loss Is Acceptable? Thresholds by Traffic TypeDefensible numbers for voice, video, and data — with the reasoning behind them.
- Measuring Loss Accurately: Why Ping Undercounts and What to UsePing's blind spots, sample-size math, and interface counters as ground truth.
- Errors, Discards, Drops: Reading Interface Counters CorrectlyWhat each counter means, which combinations point to which faults, with example output.
- Chasing Intermittent Loss: A Field WorkflowRing-buffer captures, per-hop isolation, and correlating loss with time and change logs.
- Packet Loss on Wi-Fi: Why Wireless "Loss" Is DifferentRetries versus true loss, what client statistics mean, and the metrics that reflect wireless health.
- Fixing Loss: The Remediation CatalogDuplex corrections, buffer tuning, QoS, cabling fixes, and ISP escalation — ordered by likelihood.
Quick answers about packet loss
Why am I getting packet loss, and what would cause it?
Congestion (a queue somewhere is full and dropping what it cannot hold), a bad cable or port (errors on the interface counters), a duplex mismatch, a saturated Wi-Fi channel, or a device whose CPU cannot keep up. The interface error and discard counters say which; the guides above walk the path.
How do I reduce, stop or get rid of packet loss?
Find where the packets die first: the hop whose error or discard counters rise with the loss. Then fix that hop: the cable, the duplex setting, the oversubscribed uplink, the Wi-Fi channel plan, the overloaded firewall. Loss is a symptom with a location, never a setting to turn down.
Is 1% packet loss bad? What counts as high packet loss?
On a wired network, anything sustained above 0.1% is a fault worth finding; 1% is noticeable on a voice call and bad for video, and 5% makes TCP transfers crawl. On Wi-Fi a little loss is normal and the question is whether it spikes.
How do I troubleshoot packet loss on a network?
Measure it continuously rather than once, so the intermittent kind shows up; then compare loss at each hop with that hop's interface counters. A packet train between two probes measures a whole path; interface counters tell you which device on it is dropping.
Why is my packet loss so high on Wi-Fi?
Interference, a crowded channel, too many clients on one access point, or a client at the edge of coverage retrying. The access point's retry and client counters over SNMP, graphed over a day, usually show which.
How do I prevent, detect and solve packet loss, and what does constant or spiking packet loss mean?
Detect it continuously with an ICMP sensor or a packet train, so spikes show up with a time stamp; constant loss is a bad cable, a duplex mismatch or a permanently saturated link; spiking loss is congestion at a particular time or an interfering Wi-Fi neighbour. Prevent it by fixing the hop the counters point at, not by raising a threshold.
Why am I experiencing packet loss, and is it due to the internet or my own network?
Compare loss to the gateway with loss to the ISP's next hop and loss to a public address: if the first is clean and the second is not, it is the circuit or the ISP; if the first is already lossy, it is yours. The ISP accountability series covers what to do with the evidence.
Put this into practice with SysPeep Monitor
SysPeep is network and infrastructure monitoring software for Windows. It discovers the devices on your network, polls their health and performance over SNMP and WMI, and alerts you the moment something goes wrong — all from one web console.
Download Free TrialLearn About SysPeep