Extra helpful if you can add a photodiode to the system that can adjust brightness accordingly. It costs effectively nothing in parts and should take any competent engineer less than 5 minutes to include it. Better, use multiple because redundancy is better! I wish my car had redundancy so it's entertainment panel could go back to adjusting from being visible during the day to not being blinding at night (it has brightness adjustments but that's insufficient for a car and living anywhere outside a major city)

For people doing software, press for the love of god just make that shit adjustable. Only fucking noobs hard code variables. Practicing good habits will help everyone, including you. Unless you got a serious reason not to, expose that to the users. Even if you don't think anyone will want it, I promise you, someone does. There's a lot of people and everyone thinks differently. So only lock down what needs to be locked down.

Unless you're trying to create e-waste or piss people off. Which in that case I only have two words for you and they aren't nice

To reply to this and the above comment:

Cheapest light pipe on digikey: 16 cents Cheapest photodiode on digikey: 11 cents Cheapest LED (obviously that annoying blue): 625 milli cents!

Part costs matter! Its not just the BOM, its the NRE from the increased complexity. Im not saying saying its OK, just that its inevitable considering the economic conditions.

I do board level designs and drop down LEDs. If you are not specialising in indicators, its hard to visualise how bright 10mA through your diode is going to be. Add to this that sometimes you never even see the thing you designed!

You don't even need to do that. You can measure ambient light with the LED, using it as a photodiode