Another JPLer was asking about how flexible missions can be at accommodating late breaking science discoveries or observations.

tl;dr: it can be really complicated. Here, let's walk down memory lane with @CassiniSaturn, shall we?

An April 2015 flyby of Saturn's moon Tethys showed red claw-like marks across the moon. Lots of jokes within the project (but not, sadly, in press releases) about Freddy Krueger at work. https://www.jpl.nasa.gov/news/news.php?feature=4671

What was causing them? Were they long lived, or would they disappear within months? Was there some kind of internal activity, or more outgassing from another Saturnian moon (besides Enceladus)?

The only way to tell is, of course, to grab more images when once again flying by Tethys, but flybys of most of Saturn's moons were sporadic.

By June of 2015, there was essentially ONE suitable future Tethys observational opportunity during the mission - in mid-November 2015.

Five months to plan, no problem, right? Nope.

We hadn't planned on observing Tethys back when first drawing up plans years ago for what the spacecraft would observe in November of 2015. The game plan was for a high priority dust detector observation.

So first that other instrument had to grant permission to say that they would yield their time. Which they graciously did.

Next, the planning process for each of Cassini's sequences took months, involving careful cross-checks to make sure that science observations didn't cause problems for other science or engineering systems on board the spacecraft.

One of the most difficult checks was on the spacecraft's orientation (attitude). Cassini's imaging system was pointed by rotating the entire spacecraft via its reaction wheels (think gyroscopes).

By this point in the mission the wheels were old and cranky and needed to be treated gently. The attitude engineers would run the weeks-long plan for each sequence, making sure the wheels weren't spinning too slowly or too fast

If the wheels might enter these "keep-out" zones, the attitude folks would suggest changes to the scientists, who would check that they could still obtain their observations with slightly altered spacecraft orientation.

It was a tedious, painstaking, time-consuming process, which we'd just finished. And now the icy satellites folks were asking us to redo at least some of that work, on a massively accelerated schedule.

The final sequence would be sent (radiated) to the spacecraft in mid-September, and to keep to that schedule we needed to check that the new proposed pointing would work, and we needed to do that in about a week. Lightning speed!

Everything worked out in the end -- spacecraft was fine, observations were successful. Bad weather at the Canberra DSN during the data downlink didn't ruin things.

But there's a reason why space missions don't do spontaneous.