What is 'chaos engineering' that Google is practicing for employees?



Chaos Engineering ” refers to training that enables an appropriate response when a problem actually occurs by intentionally generating a problem with a service or system. Google engineering director Dave Rangen talks about the four chaos engineering Google does for its employees.

Chaos Engineering For People Systems w / Dave Rensin of Google-YouTube

◆ Make team members randomly absent
Once a week, members who are randomly picked up from each team will work from home. Selected members will be able to perform their work, but will work from home with the rule that they should not answer any questions from other members. By doing this, Mr. Rangen says that if someone suddenly misses work, it will be a smooth training. In addition, due to work delays caused by the absence of members, it is possible to highlight whether work or information is too far away from someone.

Members who work from home can work away from other members, so there is an advantage that they can concentrate on their work without interrupting work.

Choose a person who will be the supervisor when you are absent at random. If an emergency occurs and a member who is treated as absent is necessary, the supervisor will make a decision to determine whether to contact the absent member.



◆ Deliberately delay reply
20% of team members will be selected at random, and during the work week of the week, a rule will be set that you must not reply within 1 hour of receiving an email. It is a training to make effective use of the time that the sender of the email waits for a reply, and to consider an alternative to solve the problem. In addition, Mr. Rangen says that the problems caused by delaying the response can reveal dependency relationships by department.

When delaying a reply, select a person who will be the supervisor, and in case of an urgent email, the supervisor will check and decide whether to reply immediately, ignoring the rules.



One or two people are picked up randomly every month, and when asked questions about work, we will return the wrong answer. You just have to tell a true lie, not a proper lie. Mr. Rangen says that the purpose of cultivating the ability to tell the difference between correct and incorrect answers is to convey the information of lies to the other party. In addition, it is possible to acquire the habit of confirming multiple people rather than just listening to what one person is talking about.

One supervisor is elected for each person. The supervisor will decide whether to tell the truth rather than the lie, such as when there is a problem with urgent requirements or the content of the lie.



◆ Simulate the worst
We conduct business operations assuming that major problems that are common to many departments, such as large-scale security problems, have occurred. The fact that the security failure is a simulation is limited to some people, such as the CEO and the Legal Department. If a serious problem actually occurs, it may cause a panic and delay in responding to the problem, so when the worst happens, the entire company should be able to respond calmly and ethically Rangen says.

By master1305

in Video,   Security, Posted by darkhorse_log