Management Models: A Situation Model

    In a previous article, I only superficially touched on the situational model, and it may seem that this is not a model, but an “urgent mess”. In fact, this is far from the case. Situational model - 90% consists of prepared scenarios, it remains only to compose them correctly.

    Situations are different. We all know this, the last time I cited the police department as an example, but this time I will give an example of the guard service.


    What is guard duty?
    - Corporal Karasev, the post has passed!
    - Private Kichman, post accepted!

    Everything is simple, boring and monotonous. But all for the sake of one. For the sake of an event. Here, Private Kichman saw in the dark how on the opposite mountain lights shine in the forest. According to the instructions, he wakes up his "second", shows him they count the number of bonfires in the forest on the slope. That they appreciate the situation. Kichman sends a second officer to the tent to wake up - the situation is serious, they counted 43 bonfires, and even if there are four people each, it’s 172 Czechs.

    The officer takes the only thing that is correct and consistent with the instructions and everything that he was taught the decision is to inform the headquarters and ask for reinforcements. The headquarters is only two hours from the position, and an armored personnel carrier passes it in 15 minutes.

    When the position is strengthened by the arriving and accommodated soldiers, which the officer reports in a timely manner, a nasty noise is heard in the air and the forest on the opposite slope flashes with bright lights of explosions and the noise of falling trees - this is the work of the City on the orders of the headquarters. Suppress before the attack - the same is the right decision.

    In the morning, a reconnaissance company arrives at the position and sets off on a slope to inspect the place, count the corpses and finish off the remaining ones; it might come out to interrogate someone. If you're lucky.

    So what did the solution to this situation look like in the textbook letters?
    1. In time to notice a change in situation (detect a situation)
    2. Assess the situation, the complexity and the danger of the situation.
    3. Take all measures to prevent the situation from developing and worsening the situation (call for reinforcements)
    4. Solve the situation as simple as possible (without entering into battle, apply Grad)
    5. Make sure that the situation does not happen again (send intelligence)


    We are not responsible for life and the border between Chechnya and Georgia. We solve business situations. Situations of fallen servers, sudden features and unexpected landing pages.

    It requires the same instructions, the same plan. We need to monitor the server and are regularly interested in business plans in order to immediately detect the "situation".

    We must correctly classify and evaluate it. Landing for a mobile audience? Easy! 8 hours of layout, 2 - testing, 1 - to use the registration API.

    HDD dead in RAID? And here it is clear, you need to replace the dead.

    A bug came from a live system? Yes, we need reinforcements here. Is there anyone free? “On, hold the bug!”

    We must understand the complications of the situation. Do not replace the HDD - the next time we lose data. We incorrectly estimate the time for landing - the business will lose this time, not us. Bug fix Junior? So wait three more bugs in the evening and one more in the morning.

    We must solve the situation before their complication - the spare HDD I always have is in the DC in a rack on the floor. It is necessary to replace it - I call the familiar data-center support and it changes. Of course, the baskets in the servers are numbered and the indicator blinks on a broken one - even a fool will not mix it up. For landings, there are always blanks, a run-in and documentation API for registrations - this prevents the layout of the landing from turning into the development of parts of the system.

    Unfortunately, I have very few other people's examples of successful situational management, and I don’t want to brag about it myself. If you have it - welcome in the commentary, I will be very glad to hear your examples of success in this matter.

    Still standard situations are easy to automatically find, and also automatically repair. This is a reservation, and adding space to the logical volume LVM automatically, this is Hot Spare drives in Raid.

    If you develop a list of standard situations and their solutions, your life and the life of the project will become much easier, simpler and calmer. After all, it is always easier to solve the “hellish rush” according to the proven plan than improvising on drills, especially since granite sometimes comes across poorly drilled.

    Also popular now: