I think we either overpay him or he leaves because a team will 100% offer him a 2-3 year deal. He's still at a good age, he's huge, has always played behind average to below average teams, and he has been pretty good thus far this season. (at least at 5 on 5)
I don't know if I would re-sign him.
If he keeps playing like this, he'll get something in the 2-3 year 2.5-3.5 million range on the open market with a distant chance that he gets an actual starting goalie contract similar to Koskinen's. It's a really soft goalie market.
If we trade Anderson, we'll have to gamble on a UFA goalie regardless. Better to gamble on a guy we've seen succeed in small sample size in our system already than to gamble on someone like a Reimer (after a buyout) or a Mrazek.
For reference, here is the upcoming UFA goalie market. Keep in mind, there is a reasonably high chance that all the good options here are off the market by July 1st and won't be available to us.
Lehner, Miller, Howard, and Bobrovsky aren't likely to come here, let alone be available. Lehner probably gets extended, Miller apparently wants to stay in California, I don't even have to explain Bobrovsky, and Howard will be in extremely high demand due to the lack of options for teams who need starters.
Varlamov, Elliott, and Talbot are probably the closest things to realistic alternatives to Nilsson that we might actually be able to sign. Elliott has been inconsistent the last few years, and Talbot has been awful (Although, awful in Edmonton might not count for a goalie...). Varlamov and Talbot will not be cheaper or come on a shorter term than Nilsson and there's no guarantee that either will find coming to Ottawa appealing to play behind a suspect defense when they will be trying to revive their careers (More Talbot than Varlamov).
That leaves us in a position where we can do exactly what we did with Nilsson, which would be try to identify a goalie who has been struggling else where who might do well with a fresh start and change of scenery, or whose game our goalie coaches think we can adjust to their benefit. If we find that goalie, all we're doing is kicking the can down the road another year before we're in best case scenario we're in the exact same situation we find ourselves in with Nilsson, where a goalie has done well for us over one season and we now either have to sign them to a 2-3 year contract based on a smaller sample size, or abandon that goalie.
Unless we're intent on holding on to Anderson for one year and re-assessing things completely in July 2020, the path of least resistance would be to lock up Anders Nilsson.