When you exceed specs, especially when you run more electricity through the component, you will shorten it's life. Stability depends on how far out of spec you go.
I'm not looking to see how high the overclock can go, I just want a little bit more processing power. I'll start by increasing the clock speed, then up the voltages a little at a time. If I don't see any real world improvements, I'll go back to stock.
Which -- sounds like a reasonable idea.
And ten years ago, I might have agreed entirely with Escrow4. I had come out of two careers in which it was absolutely essential that the hardware work properly and reliably. But I'd been tinkering with hardware since 1983.
So now I see my buddy Larry there, fretting over those Q9300's, and wondering himself. And VirtualLarry will remember my panic earlier this year about an intermittent instability that was difficult to replicate because of its infrequency, and difficult to assure that it was eliminated -- for the same reason. Everybody told me "It's your overclock settings! Your overclock settings, Bonzai! You've always been stingy with your voltages!"
But -- it wasn't. Nor -- had the chip degraded. Instead, the limited motherboard resources had been exceeded with "too much stuff" enabled in BIOS, like the extra HDD controllers (two) and the Asmedia USB3 feature.
What happens under this sort of scenario, where installed software, a driver or overly-ambitious hardware configuration may be the underlying cause: the user HIMSELF begins to suspect the overclocking. So you have to eliminate it as a factor while troubleshooting -- either to prove, or disprove it as a cause. This obviously complicates everything. You don't NEED the complications.
I think it's probably OK to overclock your system within reasonable voltage and thermal targets, which -- by themselves -- do not exceed specs and Intel (or AMD) expectations summarized in those specs. But like ol' Clint Eastwood said in a movie: "[A] man's gotta understand his limitations."
Next time I build a machine I intend to overclock, I'm going to determine what features I WANT TO USE, and those I don't NEED to use before I even get started. And thing is -- if you "add" something like a PCI-E card, you're going to want to validate the same settings you had before you put it in the box. IN FACT, I'd reset the system to stock before proceeding with even such a minor problem as that.
Now -- about Larry's Yorkfields. I began to worry about an E8600 system that I'd bumped up to 4.3 Ghz. I think the motherboard components were aging. The minute I realized I needed more voltage on the processor, I set it back to stock settings once and for all -- for all time. It's now no less stable than any of the most stable systems in the house.
As for the Nehalem in the OP. I think it's reasonable to expect a nominal overclock without raising the voltage much. At least find out what's left in the system with an older motherboard, keep a BIOS profile of the stock settings to which you can revert with the slightest sign of trouble. And do reliable stress tests before you pronounce it "stable."
And just an afterthought. There were two versions of Nehalem, and I'm not clear as to which the OP's belongs. If the processor is among those with a 32nm lithography, the "safe range" voltage limit was close to 1.38V. The thermal limit would be a few degrees higher than the TCASE spec -- likely to be in the low '70s C. The OP should be able to get some Mhz mileage from voltages under 1.30V, and lower temperatures than would require a top-end air-cooler -- or AiO cooler for that matter.
Coining my own Dirty Harry script-line, I like to think my own overclocking efforts as "being an outlaw within the law."
Time to speed it up a bit 
Facebook
Twitter