The scenario: a small spaceship (10-20 meters in size), or possibly a small fleet of half a dozen of them, is approaching present-day Earth from the outer solar system. We have no particular reason to be expecting it, so we're not actively trying to overcome any "stealth" technology it might be using. Assuming the people who built the ship wanted it to go unnoticed and had a slightly better tech level as our own, how close could it get to Earth before it was spotted by the public? If the answer is "not very close at all", then how advanced would their technology have to be to change that?
Minor frame challenge
Unless they have some sort of quasi-magic FTL drive, they have to be much higher up the technology ladder than we are to be able to make an interstellar trip in something 10-20m in diameter. Now, if you're using the Road Not Taken route, that's reasonable enough, but otherwise there are enormous questions of fuel, energy, and life support.
It Depends on How Long They're Willing to Take
If the sneaky spaceships are willing to take a few years to get here, they could use an orbital transfer from the gas giants to Earth orbit, and (I am not an astrophysicist) could probably position the burns so that a planet or the sun lie between the ship and the Earth when they occur. These transfer orbits are popular with people like NASA because they take very little propellant and minimize the energy required to bring a probe from location A to location B, but limiting energy expenditure also makes your spacecraft less easy to detect.
The problem is that on a mostly ballistic orbital course, getting to Earth would take years from even Jupiter's orbit.
So if your space travellers are very patient, they can arrive virtually undetected until they pass geostationary orbit. If they're not, the deceleration burns will be visible from light-minutes away, and earth-based telescopes will spot them even if we're not looking for them.
We are not always scanning the sky for threats. According to this resource, space agencies perform this task intermittently, often a few years between scans, when funding and resources are available.
When an object is detected, it's orbital parameters are determined and it is added to a catalog of Near Earth Objects. The present-day positions are estimated using this information until the next scan.
Missile defense radar systems look at objects rising into to horizon. They are not looking into the deep sky.
In 2013, the 20 meter Chelyabinsk asteroid entered Earth's atmosphere completely undetected and airburst over the Russian sky.
If your alien vehicles can perform a powered descent, and avoid big, bright re-entry heating, then the alien craft could be well on their way to landing before being detected in the sky by missile early warning systems.
What about if the fleet had the poor luck to be arriving during a scan?
According to this article, the state-of-the-art technology for detection of an 11 square foot object is 900 miles. Converting, this is about 1 foot / 0.33 meter radius, or 2 feet / 0.66 m diameter.
Scaling up to your ships of 10 or 20 meters diameter, if it's a linear relationship, they'd be detected between 13,636 and 27,272 miles. If we are scanning at the time. This altitude is the lower-end of geostationary orbit.
With our current technology, we can create materials that bend light around them, only leaving the background visible.
There is even a building started in 2020 that will have this tech applied to the whole thing. So with technology more advanced than us, it wouldn't even be a stretch to say they could apply this principle to moving ships the size of asteroids.