Google

IT Automation: It’s not that scary!

227 ratings

Course 5 of 6 in the Specialization Google IT Support Professional Certificate

This course is designed to give you an introduction to concepts related to the automatic management of infrastructure in the IT industry. Using tools like scripting languages, version control, configuration management systems, and monitoring solutions, this course will give you an introduction to automation in IT. It will have an emphasis on creating automation that’s both scalable and manageable. By the end of this course, you’ll be able to: - cite the fundamentals of programming using the Ruby language. - create basic automation scripts to perform system administration tasks. - use regular expressions to extract meaning from raw text. - manage code by using the version control system, Git. - develop and understand the benefits of tests for the scripts and automation they create. - deploy software using the Chef configuration management software and understand the principles of configuration management in the IT context. - track the health of technology systems through monitoring and become familiar with common monitoring concepts and practices Before taking this course, it's advised (but not required) that you complete the earlier courses in this program. It’s helpful to be familiar with the following concepts: - installation of software on a computer platform - basic computer network terminology and concepts - common system administrator tasks and responsibilities This course has been designed to be completed completely in-browser. You don’t have to install any software on your computer. However, supplemental readings and instructions are provided, where appropriate, if you want to install the software and tools that we discuss in order to experiment on your own systems. Except as otherwise noted, the contents on this site are licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.

From the lesson

Monitoring

In the sixth week of the course, we'll learn about system monitoring. Monitoring helps IT Support Specialists analyzes the IT infrastructure so you can identify any problems or risks in the system. We'll learn how to use one monitoring system called Prometheus and use metric visualizations to help us troubleshoot and send alerts. By the end of this module, you will set up a monitoring and alerting system, which you will need to use for your final project of the course.

Where We've Been2:32

What is monitoring?5:47

Sources of Information - Metrics4:57

Collecting Metrics6:51

Metric Visualizations4:31

Summary1:00

Meet the Instructors

Google

Now you know what kinds of metric data are available, and where you can get them.

But once you've got that data, what do you do with it?

Well, you're about to find out.

Remember that the purpose of monitoring is to get insight into the state and

behavior of the system being monitored.

One of the most common ways to do this is by visualizing the collected data with

graphs.

Graphing metric values over time can convey a lot of useful information

to the viewer very quickly, which can be a real asset during an emergency.

Some metrics are better suited to record certain kinds of data and

some types of graphs are better at visualizing specific kinds of metrics.

While it's not always the case, it's common for

monitoring graphs to represent metrics plotted over an interval of time.

The x-axis represents the time series, and

the y-axis displays the value of the metric being measured.

Like the metric types we discussed, the kinds of graphs that will be available for

use depend on the graphing tool itself.

Some of the most common ones include line graphs and heat maps.

Line graphs are the work horses of data visualization.

They're a simple, straightforward way to translate your metrics into a visual

format, and can clearly communicate the way a metric value changes over time.

Check out this one.

This graph displays the rate of queries per second a server receives

over a period of 12 hours.

One thing to check out is the shape of the graph.

The traffic increases over the course of the morning until it hits a pike at

around 8:00 AM.

After that, the number of queries slowly tapers.

This traffic pattern could be roughly correlated to the start and

progression of a typical work day.

Line graphs can be useful to get insights like this from your metric data.

On the flip side of that, Heatmaps can be useful when

you want to visualize distribution metrics like response latencies over time.

In a Heatmap, each cell represents a range of values or bucket.

The color of the cell usually indicates the proportion of values that fall

into that bucket.

In this next example, the most common values are represented by the hotter

orange and red colors, and we can see how it trends over time.

Even though both types of graphs are commonly found in monitoring systems,

they aren't the only useful visualizations that you can use to get

insight into the metrics you collect.

The documentation of each graphing platform you use should have information

about the visualization options available to you, like the Data Dog graphing primer,

or the Grafana panel documentation.

You'll find links to these resources right after this video.

Whatever kinds of graphs you use to display your metric data,

there are a few principles you should keep in mind when creating them.

First, a good goal to strive for in your visualizations is simplicity.

The best graphs can be read and interpreted easily.

That's especially important during an outage or

an emergency situation when response time is critical.

Don't forget that MTTR value.

Most graphing systems give you the option to add some kind

of metadata to your graphs, like axis labels, titles, and

a short description of what information is being displayed.

This can all help the viewer understand what they're looking at.

Second, you should have some graphs pre-built and available for viewing in

a well-known, easily accessible location, usually through some kind of URL.

Graphs should also be grouped into dashboards.

A well organized set of dashboards can help speed mitigation during an outage

by funneling the IT support specialists to the data they need to look at.

If there's a problem with the web server, having a dashboard with all

of the HTTP metric graphs together can help pinpoint the issue quickly.

Dashboards can also help in post talk analysis.

The capacity dashboard is one example,

where all of the graphs that visualize capacity information can be found and

compared so you can plan what resources you need to help with scaling.

A good rule of thumb is to create dashboards that provide answers to

the basic questions that matter to you service or system using the four

categories of monitoring signals we covered before as a guideline.