Postcard stamped in 1920 delivered 100 years later

Is this a distraction story from USPS?

Zuckerberg was quoted as saying: "What we and the other media need to start doing is preparing the American people that there is nothing illegitimate about this election taking additional hundred years to make sure all the votes are counted"

Improbable things happen frequently at scale. One possible scenario is just that this card got lost twice.

It could have slipped behind a counter at the original post office and been uncovered when that office was decommissioned 35 years later. Then it would have made its way to a newer sorting facility where either it could have been misplaced again or, more likely, been stuffed in a cabinet somewhere (as I doubt USPS was as adamant about delivering lost mail in the 1950's). That facility could easily be in operation until today - or decommissioned very recently.

No idea if anything like this actually occurred. But at 400M+ letters / day, USPS has the scale for unusual circumstances like this to be somewhat regular occurrences.

I like the scepticism!

But the fact that the story was posted by the recipient of the post card, to a Facebook Group, and then only hit the news after going viral means it probably wasn't done by a PR company.

Having done some of this "black hat PR" myself I can tell you that we don't typically cover our tracks that well.

We almost always plant bullshit by going to the news outlets themselves, or paying a blogger to put up an article on BI, Forbes, of HuffPo.

On a completely unrelated note, have you guys heard Amazon is hiring 33,000 new employees with a compensation package average of 150k? Just to help with unemployment during this time of crisis. Pretty cool of them to do that.

true, percentiles are important for telling you how your system is doing and how typical users experience it. However, I contend that you should also measure max latency, just that is is less important for understanding typical usage. But seeing your max go crazy can signal that there is something else in the stack that needs work. Silly example that I hit: max was consistently 2hr on something that p99'd at like 1min and p50'd at 2 seconds. Dug in. The 2hr was a default tcp timeout being tripped due to some connections not being properly returned to the pool. Fix the pool logic and max times came down into the few minute range. This also lowered p99.