The Soviet R-7 ICBM used 5 RD-107/RD-108 engines in its first stage in a "Quincunx" formation. Each engine had four CD nozzles, and 1 propellant pump, common for all four nozzles. The thrust of the engine is given as 1000/1020 kN at sea level. Is this thrust the cumulative thrust from all 4 nozzles, or from each individual nozzle?
Instead of seeking sources, I'll do a little bit of calculation:
1000kN per engine:
1000kN is about 100 tons-force. Wikipedia lists R7 launch mass as 280 tons. 500 tons-force of thrust (5 engines x 100 tons) applied to 280 tons of weight gives a sane, reasonable initial thrust-to-weight ratio of 1.7 which is a rather brisk, healthy pace of climb for an orbital or ballistic rocket. (compare: Falcon 9: 1.28, Soyuz: 1.4-1.6, Saturn V: 1.15, Vostok - 1.65)
1000kN per nozzle:
2000 tons-force (4 nozzles x 5 engines x 100 tons thrust) applied to the same 280 tons weight would result in launch TWR of 7.1 which is not sane. It's something expectable from interceptor missiles, not ICBMs. Japanese SS-520-5 aka Lambda 4S (with its 26kg payload capacity) comes close, but I don't know anything powered with liquid engines capable of that much TWR, never mind it's completely unreasonable and pointless unless you need to hit an ICBM that is already coming down.
Concluding: The number is given relative to engine, otherwise the majestic R7 would be zipping like an interceptor.
The nozzle is not the engine.
Each nozzle is attached to one combustion chamber, but by convention, a complex of multiple combustion chambers sharing a single turbopump, such as the RD-107/RD-108, is referred to as a single engine.
Thus the thrust figure given is for all four combustion chambers, that is, the engine as a whole. SF's sanity check confirms this.
Note that the thrust figures you give in your question are for the RD-107 and RD-107A at high altitude; at sea level the thrust for the original (R-7) RD-107 is 810kN.