Scam with spot instances

Warhammer 40,000: Darktide Gameplay Feedback
1

Background: when connecting to a game you are connecting to a server hosted by fatshark in AWS (Amazon Web Services).

There are two types of servers one can host/buy:

  • on-demand, you pay full price but once VM is acquired its yours to keep forever (SLA applies)
  • spot instances which are residue VM cores, that are leftover on physical machine (example: hpst has 64 cores, but 60 are already used by VMs, hence 4 core VM is given on discount), and they are cheaper up to 90%, but can be terminated with 2minute notice.

Guess which ones are used to host games we play on… yes, correct the spot instances. you can see that in the hostname from logs.

That explains why there is soo many connection issues at times, simply AWS gives servers to other customers who pay the full price and that causes our games to be killed.

quote from AWS docs:
As I mentioned earlier, Spot Instances represent capacity that is not in use by On-Demand Instances. When this capacity decreases, existing Spot Instances could be interrupted with two minutes of notification and then terminated.

2 Likes
2

@Reginald sruff for your next video

3

Interesting. Is there not a different scaling service or dynamic method being used? Is it all spot or is spot being used as flex?

4

Can you link is where you’ve gotten that from, because I did a 20 second Google and got something very different.

At no point in my 400 hrs across the last 6 months have I been dropped from a game that I couldn’t reconnect to on restarting the game.

"
On-Demand Instances let you pay for compute capacity by the hour or second with no long-term commitments. This frees you from the costs and complexities of planning, purchasing, and maintaining hardware and transforms what are commonly large fixed costs into much smaller variable cost.
"

Source : Amazon EC2 - Secure and resizable compute capacity – Amazon Web Services

I’d add also that there’s nothing on that page about Spot Instances terminating on 2 minutes notice?

*edit
Addition:

"
You also have the option to hibernate, stop or terminate your Spot Instances when EC2 reclaims the capacity back with two-minutes of notice.
"

Which is the customers choice rather than AWS booting you off?

Source: Amazon EC2 Spot – Save up-to 90% on On-Demand Prices

5
6

I have about 1-2 crashed servers (where I cannot reconnect) for about every 100 hours and about the same of DCs where I can reconnect. (estimate as it is so rare I don’t really count)

I’m definitely not happy about it when It happens, but based on my experience even if they do what you say, it’s working all right for me.

Admittedly I’m on CET, and playing on GamePass, so there might be a difference from you.

7

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.