Electronic – What does Expected Reliability of a Circuit Mean

If you use appropriate derating guidelines for electrolytic capacitors (keeping the ripple current in check, ensuring they stay well below their maximum temperature rating) they can last a very long time and still provide adequate performance. I've seen capacitors come out of 15 year-old equipment and still 'work' to some degree.

Fans have a limited lifespan depending on how fast you operate them. They tend to gradually fail over time.

If you have a primarily surface-mount design, consider conformal coating to help limit exposure to the elements and mitigate oxidation, etc. Products in the field for years which have some air exposure can very easily get several millimeters of dust, which makes a very effective thermal blanket.

Make sure that, if you're using RoHS parts, that you're not using parts prone to tin whiskering. (This is another area where conformal coating can be helpful).

Get the most reliable servos (and other gear) you can in the first place and over spec the servos and their attachments greatly so any mechanical stickyness can be forced out of the way.

If you are still not convinced that you have enough reliability, then look into how commercial airliners get higher reliability:

• Have multiple batteries, all of them feeding the bus though diodes, so that if one battery craps out it will not bring down the others.
• Split control surfaces and use independent servos to control each, mixed so each control surface is held below 25% authority in normal operations, that way you can count on being able to counter a stuck surface with the one left operational, just by switching off the excursion limit and accepting a lot of drag.
• Fuse each servo at the central controller, so it cannot bring down the power bus if it shorts out.
• Use servos which can report back their position and other data, like motor current like: http://www.openservo.com/ so the flight controller can detect stuck servos and try to fix the problem or at least let you know.
• Run all servos from a controller that is able to power down individual servos, so if one servo goes nuts, then it can be turned off and allowed to weather vane.
• Stick a complete autopilot with GPS and IMU in the plane (see: ArduPilot), to take it back to the launch site, if the radio craps out.

In the end I doubt any of those enhancements would do a lot of good in practice as all the redundancy will also add complexity and thus new sources of errors, just look at all the problems Airbus had getting their flight systems debugged.

... but I bet a whole lot of fun could be had building all of that redundancy.