According to the other article I read about it (https://www.forbes.com/sites/jonathanocallaghan/2019/09/02/s...) they knew about it earlier but Starlink refused to move their satelite.
Jonathan O’Callaghan published his extremely biased and inflammatory clickbait before waiting for SpaceX to present their side of the story (time from ESA tweet to Jonathan’s story was about three hours).
What actually happened is that SpaceX communicated their intention to take no action based on a much earlier estimate of extremely low collision risk. A later update from the US Air Force got lost in the mail filter, so SpaceX never read it otherwise they would have contacted ESA immediately.
Here is Loren Grush with a more mature and less inflammatory take on the topic: https://www.theverge.com/2019/9/3/20847243/spacex-starlink-s...
In time we will get a more thorough understanding of what happened, but the ESA’s fundamental point is emphasised by SpaceX’s experience: having to coordinate satellite operations through email that is handled by humans is fundamentally unworkable, and something needs to be done about automating everything before the megaconstellations take to the sky.