What is the impedance limit for using 10X probes with oscilloscopes?

A typical scope input impedance is 1Mohm + 20pF. This is not huge, even with a 100k source I get 10% error DC and 20pF is only 8k impedance at just 1MHz. A 10x porbe makes the input 10meg + 2pF, much better. It also allows for tuning of rise times by trimming a small capacitor in series with the R in the probe.

You can refer to the data sheet of the probe. Below is a typical set of specifications. Consider the 200MHz probe.

The input impedance will be 10M\$\Omega\$ || 15pF. That's a complex number in general but it will be dominated by the capacitance at high frequencies so we can pretty much ignore the real (resistive) part. At 200MHz, x10 the capacitance has an impedance of about 53\$\Omega\$, which is quite a bit of loading.

In the x1 position, it's 47pF plus the 'scope input impedance, in parallel with 1M, so more like 12\$\Omega\$, which, of course, is much worse.

You could also consider the frequency at which the capacitive reactance equals the resistance- so the magnitude of total impedance is about 7M. That's at about 1kHz- hardly high frequency but that's where the capacitive loading begins to exceed the resistive loading.

If you require very low capacitance loading then there are active probes that have only a couple pF input capacitance, but they tend to be quite expensive.