I know the current interpretation, but I also know that it's not logically sustainable.
The key here is "reasonableness", but you can't read a judge's mind, so reasonable only means what a reasonable person would infer. That is completely subjective based on an individual's implicit biases and knowledge of all surrounding circumstances.
For example, the dev team tasked with implementing this may not be familiar with the entire case, so what they consider reasonable will be different based on that. They also know much more about technology and likely have strong feelings about what is acceptable, but the judge is NOT a subject matter expert (SME) in that area. Likewise, the judge is NOT a SME on UI/UX, so the judge's interpretation of "reasonable" may well be at odds. There is also an argument that directing to these third-party sites opens 1.4B people to exploitation, so it would be reasonable to allow them access, but also to protect those users (who also aren't SMEs and often have essentially zero understanding) from potential issues by framing things appropriately. I could go on with many other qualifiers about what would be reasonable.
The judge is qualified to make rulings based on the law, but utterly unqualified to decide what the details of a reasonable implementation would be. If she were a qualified SME, she could and should have written her reasoned decision and removed the need for interpretation. If the ruling were itself reasonable and clear, there wouldn't be meaningful room for debate about reasonableness which would in turn effectively render reasonableness a mute point. As such, you can see that "reasonableness" in these kinds of cases is a crutch to protect unqualified judges making rulings they should not be making which is ironically a very unreasonable approach.
As to "good faith", who is she to dictate that you must do what she says and be happily compliant or else? After all, unhappiness will almost certainly taint all actions to a greater or lesser degree, so anything less than perfect contentment would be "bad faith" to a greater or lesser degree.
Who is the judge to single-handedly decide these people's state of mind? Has all her study of law made her an expert on what happens in other people's minds and privy to their thoughts and experiences while ridding herself of her own biases? Such a task is impossibly hard for a jury even in broad terms with lots of facts and even less possible for a single person judging the actions of many people based only on a few chat messages.
How does she ensure that every single one of those hundreds to thousands of people involved is "acting in good faith" is an impossible task and you can be assured that at least one of them at any given time isn't acting in good faith about ANYTHING. This would imply that you must ALWAYS assume that bad faith was involved.
How do you determine how much bad faith is too much? Based on outcome is the only logical answer, but in that case, you could have been explicit about the outcome in the first place and skipped all the trouble and possible excuses of misunderstanding once again rendering "bad faith" as a crutch to protect the judge from their incompetence.
She passed judgement based on the law. As long as that judgement is fulfilled to the degree specified, the state of mind of the hundreds to thousands of people involved in carrying out that judgement should not matter.
"Good faith" and "reasonableness" is a faulty and fickle metric at its best when applied to a single person by a group of peers. It is a broken and unusable metric when applied to a group or corporate entity by a single judge not an expert in the matters at hand.