Stealth is possible, but tricky. One has to account for the noise floor of space, which is not perfectly black, as much as jamming.

[MatterBeam of ToughSF did a _lot_ of work][1] exploring how to use coolant, very large wire radiators, and high expansion ratio rockets with conventional propellant to have quite effective stealth. Torch ship it is not, no, but it is very hard to detect from range. 

High expansion ratio nozzles and cryogenic coolants can also make missiles even harder to detect than the helium steamer above. One might well see their earlier stage with a sky survey, but once underway, good luck. Known ways to have low observability with radar in the present means you'd be stuck trying to find them with LIDAR sweeps or relativistic electron beam sweeps to see the scintillation returns. Stealth missiles are very doable. 

Theoretically speaking, a [Q-drive or Plasmadyne][2] using very good superconductors could be quite thermally efficient, minimizing the heat output one has to concern themselves with, though detection radiuses for a whole sky survey or a more focused scan depend on too many assumptions to give concrete answers. The linked paper uses a linac as an electric rocket, which is harder to see than many rockets, but would create RF at the least, and interactions from the particle beam and plasma would also be detectable. A stealthier option such as [nanoparticle electric rockets][3] could be even lower observable as well. As the dust (or macrons) deionizes there will be telltale but faint emissions, but not much thermal emission to worry about. Such a vessel might well want a second drive like the shuttered high expansion ratio rocket from the steamer in the first link, unless variable ISP could be developed to extend very far down to have higher thrusts, or just use it to help cool the rocket from time to time. Less conventional emissions such as whistler waves and the whole gamut of plasma waves could also be detected, but tend to be swept along with the solar wind downstream, so if you're closer to the sun than the q-drive/plasmadyne you will have a lot of trouble detecting it, unless the electric rocket is easy to detect, and it is using one. 

Jamming particularly sensitive instruments helps, though thermal emissions would need a supercontinuum laser, while looking for particular wavelengths like the spectral lines of a particular ionized element could be jammed with more normal lasers. Indeed jamming over an AU or more is conceivable with lasers in the gigawatt range, though this would mean you would have to defend your jammer. Also, I do not believe gamma emissions could be jammed at all, and x-ray jamming would be very difficult, so nuclear reactors and nuclear drives that might activate parts of the ship using either could give you away. Fortunately, if you know what your ship is made from, you can plan around your activation spectra if your setting allows for meaningful x-ray jamming. 

In all cases, full sky surveys generally detect things much closer than a focused survey would, but the sky is big, and you can't point your best telescopes everywhere at once. You're also likely better off mixing stealth with fast, "bright and loud" vessels that demand enemy attention too. 

  [1]: https://toughsf.blogspot.com/2018/04/permanent-and-perfect-stealth-in-space.html
  [2]: https://tauzero.aero/wp-content/uploads/JBIS-May-2019-Greason.pdf
  [3]: https://www.researchgate.net/publication/231125597_Nanoparticles_for_solid_rocket_propulsion