In my opinion, the Rigols are too close to similar Agilent/Fluke models in price to be worth it. Those brands are known for good quality, Rigol is quite unknown. Their scopes are wildly popular because Agilent re-branded them and they were one of the cheapest scopes that doesn't suck, but that doesn't mean they're great (it's just hard to do better without paying more). The DMM's are fairly unknown as far as I know. Documentation and support usually seems to suck. Do they publish at least full calibration documentation for their bench DMM's? They don't for their scopes, which is stupid in my opinion.
If I was buying new, I'd be looking at the Agilent 344* stuff (eg. 34405A, 34401A) and the similar Fluke stuff. I've heard good things about Picotech, but don't think they're much cheaper.
For hobby use, a used meter is much better value in my opinion. You can even get them calibrated from some ebay sellers. I would ask for a sample certificate and verify if they actually have adequate equipment that's in cal, or if they just plug it into a voltage standard and verify if it displays something close to 10V. This only makes sense with the more expensive units, since calibration is more-or-less a fixed cost, it doesn't make much sense to pay a premium to get a <$100 meter calibrated in my opinion.
5.5 digits is fine for almost any use. 4.5 digit is usually good enough, but is not significantly better than many hand held meters. Note that bench meters (especially the older ones) usually focus on doing a few measurements (typically at least DCV and resistance, often ACV, sometimes DCI/ACI) very well, so they don't usually offer features like diode tests, continuity or capacitance. The newer models usually have more bells and whistles. Since they're often designed for system use (in automated test setups), they can usually do many readings per second, especially the newer ones. This often translates in quicker response, and makes it easier to see that a signal is unstable. I often get annoyed at the low update rates of my Fluke hand held meters when using them side-by-side.
Some attractive units in my opinion (in no particular order):
Keithley 192/195A/196/199: I like their LED displays. The latter two are more recent and have a nicer interface. The 192 is really limited, no current, and even ohms and AC are optional. Most of them have GPIB standard. The 199 has weird current ranges (either 3A or 30mA). The 195A's accuracy is not great for a 5.5 digit unit. The 196 is basically the 6.5 digit version of the 199, with a few extra features and slightly nice controls. The 192/195A/199 can occasionally found for around $100, the 196 is more expensive.
Fluke 8840/8842: Nice units, but usually overpriced in my opinion. They usually sell for $200, but have similar features to the 5.5 digit Keithley/HP meters, which are usually much cheaper. The VFD is nice, if it's not dim from being on 24/7 for many years. AC measurement is optional. I think GPIB is also an option.
HP 3468A/3478A: They are quite similar, the 3468A lacks GPIB and one lower range, and is usually significantly cheaper. I dislike their unlit LCD displays, and they lack features like relative measurement, but are fine units.
HP/Agilent 34401A: This is still a current model, and is significantly more expensive than the others in this list. Nice VFD, nice controls, 6.5 digit, lots of features. Has continuity and diode tests. I think it usually costs like $400 or so. Has GPIB and RS-232.
Keithley 2000: Competitor for the HP 34401A, specs and features almost identical. Fine unit. Not as popular, I would want to pay slightly less for it than the HP 34401A in similar condition. Has GPIB and RS-232.
HP 3456A/3457A: Huge 19" reference meters. 6.5 digit, good accuracy (slightly better than the 34401A). These are sometimes available for cheap, but shipping can be expensive. The 3456A lacks current ranges, but has a nice LED display. The 3457A has the same bad unlit LCD display as the 3468A/3478A, but is a newer unit, and is faster for system use. Has some more math features, will do 7.5 digit via GPIB. Not very important features in my opinion. Some people claim that the 3456A is more stable. Both have resistance ranges into the Gohms and pretty low ranges (good resolution).
There are some other brands like Datron, Solartron, Prema and Schlumberger which made precision DMM's similar to the HP 3456A/3457A, many of them are fine units, but they are less popular, and documentation is often harder to find (both Agilent and Keithley make documentation for their older stuff available on their website).
There are also various older and less capable units, like the Keithley 'portable' series (eg. 197(A), 175(A)) and the Fluke 8800/8810, but if you're considering new Rigol units, you probably want more features.