Latency Monitors

Latency Monitors (previously called Connection Quality Monitors) measure round-trip latency, jitter, and packet loss to hosts on the internet. They appear in the Monitors section of the main window.

What they measure

Metric	Description
Latency	Round-trip time for a single ping, in milliseconds.
Jitter	Variation in latency over time. High jitter suggests an unstable connection.
Packet loss	Percentage of pings that didn’t get a reply.

Higher latency, jitter, or packet loss can indicate problems on your network or along the route to a host.

Single-hop vs differential

A Latency Monitor can target a single host (for example, 1.1.1.1) or measure the difference between two hops (for example, Router → ISP). Differential measurements isolate problems to specific portions of the path, which is how the Internet Dashboard measures latency to your router and ISP.

A differential monitor shows an estimate, and never reads 0.0 ms. A difference under 2 ms reads <2 ms. When the monitor can’t measure the difference, a caption beside the latency explains the value:

• Whole path — the value is the latency of the whole path to the far host, because the monitor can’t measure the hop between the two ends.
• Slow ISP hop — your ISP’s router answers pings more slowly than the host behind it, so the monitor can’t measure the difference.

Hover over the latency to see the caption when the row is too narrow for it.

Section header

Latency Monitors share the header of the Monitors section with your other monitors. Its Shared Y-Axis Scale → Latency puts every latency monitor on one scale, so you can compare latency directly.

A latency monitor can also be one of the Internet sites that make the Internet status of the dashboard. See Use a latency monitor as a site.

Reading the graph

Each monitor shows the latest latency value, a colored quality indicator, and a graph of recent measurements. Hover over an expanded graph to reveal the graph toolbar below.

A single monitor packs more detail than it first appears:

Along the top of each monitor, from left to right:

• Disclosure triangle — expands or collapses the monitor’s graph, leaving just the title and current latency visible.
• Status light — a colored dot showing the monitor’s current state: green for Active, yellow for Warning, red for Error, and grey for Disabled.
• Symbol — the icon representing the target or path being measured. This can be configured in Settings → Latency Monitor.
• Name — the monitor’s label. Set this in Settings → Latency Monitor.
• Current latency — the most recent round-trip time, in milliseconds.

The graph itself adds a couple more details:

• Latency graph — plots recent round-trip times, colored by quality so spikes and slow responses stand out at a glance.
• Timeout / No response — pings that didn’t get a reply are drawn as red marks, making packet loss easy to spot.

A monitor that gets no result for 3 minutes shows an error instead of Starting…: No response. A firewall or this network may block PeakHour.

Configuring

Add, remove, and customize Latency Monitors from Settings → Latency Monitor, where you can adjust polling intervals, latency thresholds, packet count, and colors.

Graph toolbar

Hover over an expanded graph to reveal its toolbar:

From left to right:

Control	Description
Probe-method pill	Appears only when the monitor has fallen back from standard ping. The label shows the method in use — UDP TTL, TCP, or HTTP. See Probe methods below.
Zoom out	Zooms the graph out, showing more time.
Zoom in	Zooms the graph in, showing less time.
Share	Shares an image of the graph via Mail, Messages, AirDrop, etc. The image has the monitor name, a legend with the current values, the graph, a scale and a time axis.
History	Opens the History view for this target.
Settings	Opens the settings for this monitor in Settings → Latency Monitor.

Probe methods

Latency Monitors normally measure round-trip time with standard ICMP ping. If ping is blocked or stops getting replies, PeakHour automatically falls back to another method so monitoring keeps working — stepping down through UDP TTL, then TCP, then HTTP until something succeeds. For a site on the internet, the HTTP method uses a secure HTTPS request. Any reply from the site counts, also an error page or a certificate error, so the latency is the time to that reply. A host on your local network keeps plain HTTP.

When a monitor gets no reply, its status in Settings → Monitors gives a short reason, for example No response, Connection failed or Unexpected response.

Whenever a monitor is running on anything other than ping, the toolbar shows a small pill with the active method’s name. PeakHour keeps retrying the better methods in the background and clears the pill automatically once ping recovers (usually within a few minutes).

Note

The pill is informational — it tells you a connection is being measured by a fallback method, not that anything is wrong. You don’t need to do anything; it disappears on its own when standard ping starts working again.