Add error signal offset graph

[adriangudas]

For: https://github.com/Shopify/resiliency/issues/6611

This adds a way to see the difference between the ideal error rate (as calculated by the PID controller) and the actual error rate that the user requested as part of the experiment.

Open questions around this are:

• How can we use this to get a deeper understanding? We're interested in fluctuations in this value, so maybe we can graph that (the derivative) and see if we can get some information from that. Also:
• We may need additional graphs/experiments to try to highlight the effect of elevated error rates on the ideal error rate
• How does a 30-minute incident affect the calculation of the overall trend? Ideally, the calculation should not be skewed by this.