Measuring happiness

 

Congratulations! You passed!

Grade received 100%

To pass 100% or higher

1.

Question 1

What do you think makes a metric a good measure of happiness? Choose all responses that are correct (2 correct responses).

1 / 1 point

It is derived by scraping social media for complaints

It closely tracks well-known system behavior

It shows long-term trends clearly

Correct

Correct! A smoothed-out signal over a long time horizon is more valuable for an SLI, because these metrics tend to have lower variance and higher signal-to-noise ratio.

It reacts quickly to changes and outages

It is a close approximation of the user experience

Correct

That's right! We want to measure the user experience as directly as possible. We assume that when that experience is poor enough, we are not meeting our users' expectations of service reliability and they are therefore unhappy with our service.

Increasing reliability

 

Congratulations! You passed!

Grade received 66.67%

To pass 66% or higher

1.

Question 1

When improving service reliability, what are some ways you can improve the time-to-detect? Select all that apply (2 correct responses).

1 / 1 point

Develop a playbook for outage scenarios.

Implement monitoring for your service.

Correct

Great job! Implementing monitoring and automated alerting are some examples of the ways you can improve time-to-detect.

Add replication to your service regions.

Add automated alerting that pages a human.

Correct

Great job! Implementing monitoring and automated alerting are some examples of the ways you can improve time-to-detect.

2.

Question 2

When improving service reliability, what are some ways you can improve the time-to-resolution? Select all that apply (2 correct responses).

1 / 1 point

Create automation for manual mitigation steps

Correct

Correct! Automating outage mitigation steps, such as draining from one zone to another, can help reduce the time-to-resolution.

Running in multiple zones so that failures are less likely to happen.

Add automated alerting that pages a human.

Develop a playbook for outage scenarios.

Correct

Great job! Developing a playbook can reduce time-to-resolution as it can help humans dealing with an incident troubleshoot and root-cause more effectively.

3.

Question 3

What are some operational ways you can increase reliability? Select all that apply (3 correct responses).

0 / 1 point

Periodically collect data on poor reliability regions and focus extra efforts on those regions.

Correct

Great job! Those are some examples of ways to improve reliability from an operational standpoint.

Providing post-mortems

Correct

Great job! Those are some examples of ways to improve reliability from an operational standpoint.

Creating an operations center where incident responders can watch graphs

This should not be selected

NOCs are not a great way of increasing reliability: they rely on humans instead of automation and they tend to encourage a focus on symptoms rather than root causes.

Work with the development team to provide input on the feasibility of achieving targets.