Chiming in here -- some metrics are B.S. metrics (like how much money you saved company, I gloss over these myself), but some are valuable. For example: "designed and developed system Foo which ran in production for X years and processed Y events per second utilizing Z nodes."

I agree. Some projects have SLAs. It could be beneficial to mention that those were met or expectations were exceeded.

In some cases saving a company that much money is very believable. For instance if you fixed a transaction processing bug at a company who does $84K a month and it lasted a year, you saved them a million dollars.