FWIW: From years of cooking on open fires in the bush, it is not an instant to go from candle to conflaguration. Under exceptionally dark nights, you can see the flare of a match from a long distance. At night it may be sufficient to have a dark lantern, and open the shutter of it. I would normally think of making a pilot fire (candle or lamp) into a easily visible at 15 miles fire to take several minutes.
In day it is somewhat easier, as a fire kept burning can be turned into a column of smoke very quickly by dumping green vegetation, wet grass or straw on it.
The Armada is far more likely to be seen in daylight. No sensible captain wants to be close to a not very well known shore in darkness, so if at all possible they would have been well out at sea -- you would want to be further from shore than your maximum likely error dead reckoning. I would hazard a guess of at least 15 miles off shore.
So the night transmission problem would only occur if the the fleet was spotted late in the day, near or after sunset. (Right after sunset, the ship's rigging would be silhouetted against a bright sky.)
Using heights would be essential. At 15 miles, there is about a 200 foot drop in the horizon. You would need to around 100 feet above sea level to even see the rigging tops. A sight path that goes near the ocean surface is badly distorted by changes in temperature and air density.
When I was canoeing in Canada's north, from eyes 3 feet above water level I could distinguish trees as being individual trees up to about 3 miles. At 6-8 miles headlands became silhouettes. And around 10-12 miles you'd get mirage effects, and only the tallest land forms would stand out. Now, nearly all my visual path was in the disturbed layer, but it illustrates the problem.
In either night or day transmission, you need a visibly large contrasting signal to be absolutely clear and unambiguous.