5

I've found something extremely werid on my EC2 instance today; a peak on the "Network In" that goes beyond anything that had happened before, as you can see in the screenshot. Any idea about how to check some logs that help me understand what has caused this peak???

The apache access logs are normal,no more petitions than normally...

Screenshot:

enter image description here

MadHatter
  • 79,770
  • 20
  • 184
  • 232
vanderflo
  • 71
  • 4
  • 3
    a) you can upload images directly to Server Fault and we won't be bombarded with adds. b) Supposedly the advantage of a trained human operator is that they can recognise anomalous data as erroneous and discard it. It is extremely unlikely that you really had a peak network speed of 3.5 exabit per second... – HBruijn Aug 23 '14 at 22:02
  • Hi, I couldn't upload the image because I didn't have enough SF reputation :(. And yes, as you say, it's unlikely that this peak happened... but the DB collapsed as it never did before, so something weird happened at that time for sure, just wanted to check if any other person has experienced something similar or there's a way to see what happened... – vanderflo Aug 23 '14 at 22:13
  • 2
    That's nearly a billion gigabytes per minute, so if nothing else I'd contact AWS support with an "uh..." – ceejayoz Aug 24 '14 at 00:54

1 Answers1

7

This is clearly a bug in the software computing the traffic rate. You need to show the image to the vendor of the software, which produced the graph. That's the only way, you are going to get an answer as to what happened.

Probably some negative number showed up where a positive number was expected, and wrapped around to around 18EB. Use that erroneous number in further calculations, and graphs like this happen.

You can not expect to find an explanation on the server itself, because there is no way this amount of traffic could have happened on the server.

kasperd
  • 30,455
  • 17
  • 76
  • 124