Thanks, but I was actually wondering more about the apparent time of the flare (say, peak brightness since as you say the duration changes) based on the observer's speed.

Say we're both on Earth 26,000 years ago. You stay on Earth and I shoot off at C/2 in a direction away from the SMBH. You wait 26,000 years and then see the flare. How long do I wait in my own reference frame before I see the flare?

My guess intuitively is that time dilation and the extra distance cancel out and I still wait exactly 26,000 years (no matter how fast I go), but I'd have to dig out my old physics textbook to figure it out properly. Also, what happens if instead I shoot off towards the SMBH? Or at 90 degrees? I'd obviously see the flare sooner in my reference frame, but how much?

It seems like there's a general rule that, by moving, you can decrease the time until you see events (at least those communicated by light) but you can never increase it.

The Lorentz factor here is 1/sqrt(3/4), around 15% over unity. So the elapsed time for you to see the flare is 2/1.15 * 26ky, or 45 ky. In general, for any λ multiple of c, the elapsed time multiple is sqrt(1 - λ^2) / (1 - λ), which reduces to 1 + λ in the λ << 1 case. You can try plotting it for λ from -1 (c directly towards the event) to +1 (c directly away from the event) to get a feel for how it behaves.

That's great, thanks! For anyone else who wanted to see the graph, here's mine: https://www.desmos.com/calculator/kihnh5ohrv

If you (somehow) keep accelerating in one direction, there are events you will never observe. It's called a Rindler horizon. Your world-line traces a hyperbola... which will always outpace photons from certain spacetime locations.

https://youtu.be/7cj6oiFDEXc