Made some substantial progress on this project. Here's a what I've found:
In order for the compressor to operate correctly you HAVE to make the discrete oscillator. The interaction of the original blocking oscillator with the rest of the circuit is quite large and there's no way you can just drop in a 555 circuit and hope it will work out fine. With discrete oscillator the circuit requires almost no tuning, all ratios and the threshold are correct. Threshold seems to be in dBV, not dBu.
For those who chose the 555 chip as an oscillator. Check your ratios! You may have to fine tune the PWM mark-space ratio to get the ratios correct. The threshold will probably be off as well.
Now, here's the curious problem that I have both with the 555 and discrete oscillators. When the compression kicks in, there's a strange looking waveform, almost as if some low frequency modulation is going on. This only happens when the compressor goes from no compression to compression (when the input level increases). I'm including the picture of the waveform. You can see that there are two places where compressor starts to work. They both look different, but conditions were absolutely the same.
I wonder what other, brighter then me, members of this forum can suggest.