Imagine you have a virtual machine. Its running peacefully inside the cloud somewhere. You are a bit curious how close it is to max capacity. You log in and run top. And then you are baffled. Maybe it shows 99% but you think it can do more. Maybe it shows 1% but you are skeptical it has 100x to go.

But wait, what is that field @ the far right in top? %st?

That field is called 'steal time'. And steal time works like this... Your 'top' uses standard unix accounting. And, it expects time to sum to 100% of (idle/system/user). So when you are running inside a shared infrastructure, there can be a fourth, which is, some other virtual machine 'stole' some time. This is an involuntary wait, e.g. it wanted to be in system/user, but the ticker wasn't available.

Another way to look @ it... if steal time is quite large, then you are running on a cloud instance w/ a 'noisy neighbour' and maybe a bit more oversubscribed than you want.

So you have a couple of cases:

  1. Top is showing you are pretty busy, but steal-time is low... You need to buy more VCPU (you are busy but the host is not)
  2. Top is showing you are not that busy, and steal time is high (you need to run on a less oversubscribed cloud)

There's a neat article here w/ some graphs for those interested in this being empirically plumbed out.

Last evening I had a fairly crappy flight  (much delays, and i'm in seat 53A, back left latrine row non-recline). And of course, someone else was also assigned seat 53A.

At any rate, the inflight entertainment system is kind of new looking and, what is that, an Ethernet jack (and PS/2 keyboard jack!). So... If i had brought along my PS/2 model 80 I could have used its keyboard and plugged in some Ethernet to it??? Weird.

As part of the fairly crappy flight from the previous flight, we had many hours to enjoy the delights of the ankara airport. Namely, a cafeteria-style food place, sort of like what you had in your university residence.

As is my normal MO, when i can't read anything, i just pick what looks interesting. So I pick this drink because it looked 'all good' in the name.  its called Salgam Suyu. My first clue something might be up is when the local said "Whoa, you got that!" and then described it as a 'spicy pickle juice'. Well, its sort of the flavour of kimchi/rotten sauerkraut (and a cup of salt), w/ the texture of a carrot soup. It also has an odour best described as feet. I linked a youtube video below where someone reviewed it.

well, one sip later, i'm like, it can't be that bad, maybe the first sip was an abberation. Nope. I managed to ingest 5mL, which is likely near the LD50.

It been a bit damp here the last couple of days. And, it appears, Sam Knows whitebox probe detects that! if you look @ this chart, the July-17th 4am drop is where the rain started, and the 12pm is where it stopped. It then started again around 6pm, and stopped around 1am. And then again around 4am today. I'm on HFC, which is normally pretty water repellent. Next to see if it can detect mosquitos and other summer events!

So my Sam Knows whitebox probe has been running for a couple of days now. And here is the UDP latency graph. Notice something weird? It is bi-modal.

So why could this be? Every 2nd test is in a different band of latency? I did a 1 day capture of the UDP traffic from the probe. I haven't looked yet, but, are there 2 servers? Is there a 2-port lag to it and one port is busier than the other? I also ran smokeping on a home machine to the same server they use, it reports as follows, so i guess there is some support for a range.

Hmmm... The average will be precisely where none of the packets lie 🙂