An instrumentation amplifier amplifies the small difference between two inputs while rejecting voltages common to both. That makes it the standard choice for strain gauges, thermocouples and ECG electrodes.
The three-op-amp design. Two input buffers share a gain resistor R_G, followed by a difference amplifier. With equal resistors R in the first stage and a unity-gain difference stage:
G = 1 + 2R ÷ R_G
With R = 10 kΩ and R_G = 1 kΩ: G = 1 + 20,000 ÷ 1,000 = 21.
Changing one resistor, R_G, sets the gain without upsetting the balance of the circuit. That single-resistor adjustment is the reason the design exists.
Single-chip versions give the gain formula on the datasheet. A common part uses G = 1 + 49.4 kΩ ÷ R_G, so a gain of 100 needs R_G = 49.4 kΩ ÷ 99 = 499 Ω, which is a standard 1% value.
Why not use one op-amp? A single difference amplifier has low input resistance and needs four matched resistors to keep a high CMRR. The instrumentation amplifier has very high input resistance on both inputs and keeps its CMRR when the gain is changed.