This flashing, which you say you did yourself?
View attachment 371360
There is the red coloured bit; to the left of that there is something else, which I imagine is what it was all like before you did your repair: what is it exactly?
Generally in this kind of situation, ideally there would be the roofing (felt in this case) and where it meets the wall it would bend upwards and tuck behind some proper flashing, most likely lead. That lead flashing should go into a kind of deep slot in the brickwork behind the render. So any water running down the surface of the render, or even behind the render, gets directed outwards, down the outside of the lead flashing, then onto the roofing and away.
The alternative is some kind of drip detail at the bottom of the render; it might be a kind of metal angle a bit like a plasterer's stop bead, again the idea would be that it would catch any water running down the surface of the wall and shed it away from the wall, and not let it get near any kind of top edge of the felt roofing where it's folded up the wall.
The "repair" job doesn't do either of these things - it looks like the felt is folded up the wall and probably not really attached to it properly, then there is a kind of a messy bottom edge of the render, then the gap inbetween is filled with some kind of mastic or something. This is pretty much bound to fail, at least after a while, because even though they might not be obvious there are probably lots of tiny gaps between the render and the mastic, and the felt and the mastic, and water can find its way along these, either by capilliary action or driven by wind.
You say the problem occurs when it's windy - never underestimate the power of the wind to force rainwater uphill, through tiny cracks and so on.
Therefore, to me it's plausible that on a window day, water gets blasted up the slope of the roof and finds it way in through the dodgy repair job.
Like
WouldBe says, once water gets under a felt roof it can track along quite a way and you might see damp appearing quite distant from the place the water's actually getting in.
I might be wrong, can't say for sure this junction is where the problem is, but it's one of the things I'd be taking a look at.