***: Lucifer333 has joined #arpnetworks mkb: if it can't handle leap seconds it probably can't handle daylight savings time either brycec: On the contrary, DST is essentially ignored since unix-like systems track time as seconds from the epoch (Jan 1, 1970 UTC) -- DST is essentially nothing more than formatting for dates. (Yes, log ingestion may be affected depending on date formats, etc)
Leap second on the other hand is that very same epoch seconds counter repeating itself for 1 second (simply put). This kind of difference is like saying "my house is painted blue half the year" versus "the universe stopped existing for a moment."
So, programs that do closely track and tie system time (the clock crystal ticking away time in your computer) with UTC are going to have an issue because it will suddenly lose sync.
Most programs though, they just call gettime(), strftime(), and that's it.
Final numbers for yesterday on zeit.arp: 12/16/2016 9.63 GB | 9.47 GB | 19.10 GB | 1.85 Mbit/s
And as it turns out, it's overflowed the conntrack table so that may be why we're not seeing 20mbps... ***: trobotham has quit IRC (Ping timeout: 246 seconds)
trobotham has joined #arpnetworks
ziyourenxiang has joined #arpnetworks
carvite has quit IRC (Ping timeout: 245 seconds) brycec: (heh... Just noticed/realized that I've been rate-limiting NTP traffic all along too, 10conn/second + 10pps)
Also I've been seeing a decline in traffic over the last 4 hours or so. What had been >80k unique sources is currently 36k and falling overall
Could absolutely just be temporary, I've seen other lulls like this before too. https://www.dropbox.com/s/421ubrqu20fda1q/screenshot_2016-12-17_16-04-21.png?dl=0# BryceBot: Dropbox photo: "https://www.dropbox.com/s/421ubrqu20fda1q/screenshot_2016-12-17_16-04-21.png?dl=0#" brycec: (aaand there's a spike. *sigh* Too much to hope for, apparently.)
(167k) nathani: 10conn/sec across all clients or per Ip? ***: Lucifer333 has quit IRC (Quit: Leaving)
ziyourenxiang has quit IRC (Quit: Leaving)