There are time-based "triggers" and there are performance-based "triggers."
I myself use a time-based trigger. I rebalance every 12 months. Bill Bernstein's research in The Intelligent Asset Allocator pointed to 18 months as the ideal time based on backtesting, if I recall correctly. More frequently than that, and you lose the tendency of "hot" asset classes to remain hot in the short term; longer than that and you gain the tendency of asset classes to "mean revert" to the average over time.
These days trading costs are negligible. On a $100,000 portfolio with 10 asset classes and $10 trades, performing 10 trades a year (assuming you used stocks or ETFs and not mutual funds), you are incurring just 0.1% in expenses annually. And many times your allocation will remain close enough in some asset classes that you don't need to rebalance all 10.
As a rule of thumb, if rebalancing will cost you more than about 0.5%, it's probably not worth it unless your allocation is WAY out of whack since rebalancing the right way probably adds little more than 0.5% to total return over strict buy and hold.