Proton decay has not been observed, but even if it happens, it needn't be an obstacle to life, as such. For humans in anything remotely like our present form you need protons, but not for life in general. Entropy, however, is a problem. All life depends on having an energy gradient of some form or other; in our case, basically the difference between the temperature of the Sun and that of interstellar space. Now, second thermo can be stated as "All energy gradients decrease over a sufficiently long time"; so eventually, for any given form of life, the gradient it works off is no longer sharp enough to support it. However, what you can do is to constantly redesign life so that it will be able to live off the gradients that will exist in the next epoch. You would be trying to run the amount and speed of life down on an asymptotic curve that was nevertheless just slightly faster than the curve towards total entropy. At every epoch you would be shedding life and complexity; your civilisation (or ecology) would be growing constantly smaller, which is of course a rather alien thing for twenty-first century Westerners to consider. However, the idea is that by growing constantly smaller you never hit the wall where the gradient just cannot support your current complexity anymore, and instantly collapse to zero. An asymptote that never hits zero is, presumably, better than a curve of any shape that hits the wall and crashes - at least this is true if your goal is longevity; of course, pure survival is not the only goal of humans, so there's a value judgement to be made there. You might decide that it's better not to throw anyone out of the lifeboat and all starve together, rather than keep going at the price of endless sacrifice and endless shrinking. And, of course, if we can extrapolate to such incredibly distant beings at all, there's going to be quarrels over exactly who gets thrown out, and the resulting conflict might well make the asymptote shrink drastically, or collapse, as resources are used to fight instead of survive. To survive literally forever you need to be lucky every time; entropy only needs to be lucky once.

That said, even with total entropy you get the occasional quantum fluctuation that creates a small, local gradient again - in fact, arbitrarily large gradients if you wait arbitrarily long times; if somehow you were able to survive the period between such events, you could indeed live for ever. In fact, if you are able to wait long enough you will see a quantum fluctuation the size of the Big Bang. The problem is, of course, that a human, and probably life more generally as well, is extremely low-entropy compared to the sort of universe you get at 10^1000 years. In fact, interstellar space from our era would look rather low-entropy compared to that stuff. So the difficulty is to protect yourself against the, as it were, sucking vacuum that tries to rip the low entropy out of your body, without using up your reserves of energy on self-repair.

Overall, I'd say it doesn't look utterly hopeless, although it is subject to a Fermi paradox: If survival over arbitrary timescales is possible, why don't we see any survivors from previous BB-level events? If my account is correct, it seems unlikely that ours is the first such fluctuation.