just now

Preempting System Issues

When major system issues occur, it's already too late. With large scale problems, it usually takes a while to diagnose the issue, fix it, get it through testing, and get a fix deployed. This is true even of hotfixes.

Listen to this episode

0:00 / 0:00

Summary

Simple systems fail simply. Complex systems also fail simply, but their interconnectedness with other systems makes mitigating failures much more complex. Past a certain level of complexity, system failures are an emergent property of the system – that is, the set of system parts has a set of failure cases that the individual parts do not have by themselves. This means that it is more difficult to predict what can go wrong with a system. At some level, prediction is nearly impossible. However, you can predict many of the things that are likely to cause problems, simply by engaging in a few fairly simple thought exercises, you can greatly reduce the number of unexpected problems that your system encounters.While it can be tempting to wait until a problem occurs to try to mitigate it, this is unwise in a production system that other people are dependent on. A system failure usually costs money at a minimum, and the problems can be far more severe than that. As a result, it's common for software services to include a Service Level Agreement or SLA, that dictates expectations about the frequency of system outages, response times, and time expected to complete work. Even if your system is engineered so that it doesn't completely fall over when a problem occurs, it can still violate an SLA and cost money. The consumers of your application probably have their own clients who have their own expectations. SLAs tend to bleed inward from clients to the services that they use and then to the services that those services use.In contrast to SLAs, systemic problems, including both errors and latency tend to bleed outward from one service to its clients and then to the clients of that service. As a result, when you are thinking about how to find potential systemic problems, it's often best to think of these problems from two different angles. That is, you need to consider how errors and latency will bleed out as a result of a problem, while also considering how SLAs bleed in to put more stringent expectations on your system than you might expect. In effect, you are dealing with a balance between tolerance for errors and difficulty in error mitigation. Depending on how critical your system is to your clients, these expectations will vary.You can't prevent every problem in a system, but you can usually prevent a large percentage of them by planning ahead. However, until you've encountered enough unexpected problems, it can be difficult to envision how something can go wrong, or even have a realistic thought process for thinking about what can go wrong. However, if you go through the thought exercises we've outlined here, then you have a good chance of preventing most of the problems that will plague a complicated application. While this doesn't fix everything, it can give you enough breathing room to fix the truly unusual problems that you'll occasionally encounter.LinksJoin Us On PatreonLevel Up Financial Planning The post Preempting System Issues appeared first on Complete Developer Podcast. Hosted on Acast. See acast.com/privacy for more information.

First published

07/13/2023

Genres

technology business careers

Duration

40 minutes

Parent Podcast

Complete Developer Podcast

View Podcast

Share this episode

Similar Episodes

No similar episodes found

Similar Podcasts

  • Think Big Property

    08/06/2020

    Tyrone Shum and Nhan Nguyen

    The Think Big Property Podcast is where Nhan Nguyen earns millions from property development and Tyrone Shum has millions of questions.Nhan Nguyen is a successful property developer with years and years of experience in the property industry. With over 100 deals completed throughout his property career, there is no one better to talk to that can provide us with some of his expert advice on how to make money in the property world. Tyrone Shum is the host of Australia's No.1 Real Estate Podcast: Property Investory receiving over 1.6 million downloads and over 200 interviews of property investors and expert stories.Together, they bring a wealth of knowledge and share their experiences from real-world examples, case studies and stories in property investments. Hosted on Acast. See acast.com/privacy for more information.

    Clean
  • SuperPotion™ | Branding & Marketing des marques de boissons, bières et spiritueux — Avec Ludovic Mornand

    08/06/2020

    Studio Blackthorns

    SuperPotion™ est un podcast destiné à aider les entrepreneur·e·s, startupers, créateurs et créatrices de la filière Boissons, Bières, Vins & Spiritueux. 🔮⚗️🌕→ Site officiel (https://superpotion.fr) → Ebook "Lance ta SuperPotion avec succès" (https://superpotion.fr/lance-ta-superpotion-avec-succes-un-guide-branding-complet-pour-developper-consciencieusement-ta-marque-de-boisson/) → Plus de 160 packaging designs prêts à l’emploi pour ton prototype ou ta nouvelle boisson (https://superpotion.fr/packaging-design-concept/)→ Besoin d’un déblocage ou d’un avis extérieur dans ton projet boisson ? (https://app.harmonizely.com/studioblackthorns/potionultra) ++++++📝🎤🎛️ Conçu, écrit et réalisé par Studio Blackthorns.→ Site Internet : https://blackthornsdesign.com/Hébergé par Ausha. Visitez ausha.co/fr/politique-de-confidentialite pour plus d'informations.

    Clean
  • MLM Marketing, Training, and Lead Generation | MLM Complete Package

    08/06/2020

    Chris Lopez & Jason Wells

    The MLM Complete Package podcast is all about helping you develop better communication, sales, marketing and leadership skills that translate into business growth by becoming a Complete Networker.

    Clean
  • Pawsey Capital Refresh

    08/12/2020

    Pawsey Supercomputing Centre

    Welcome to the Pawsey Capital Refresh Podcast. The purpose of this podcast is to:1. Communicate updates and changes to the procurement of Pawsey’s $70 million capital refresh from 2019-2022 (projects completion)2. Remain transparent and accountable in Pawsey’s use of this government-granted funding3. Develop a platform that can be used as a call to action for feedback on the projectThe team aim to release an episode every six weeks with an update from the appropriate procurement lead.We also encourage you to complete a survey after listening to the podcast, to ensure the team are providing you with the information you want to hear.https://www.surveymonkey.com/r/Capital-Refresh-Podcast

    Clean
  • David Hubel

    08/13/2020

    David H. Hubel received the 1981 Nobel Prize in Medicine for his discoveries concerning information processing in the human visual system. His work shed light on one of the mind's darkest mysteries, the process by which the brain interprets the signals it receives from the eyes. Born and raised in Canada, he attended McGill University Medical School and completed his residency in neurology at Johns Hopkins University in Baltimore. He began his research in the visual cortex of the brain while serving in the U.S. Army at Walter Reed hospital, where he developed the microelectrode and the modern hydraulic microdrive to observe neural activity in the visual cortex. He joined the faculty of Harvard Medical School in 1959. His pioneering studies of the process by which the brain perceives contour, motion, depth and color have enabled treatment and prevention of numerous visual impairments in adults and children. In this podcast, recorded during his appearance before the Academy of Achievemen

    Clean
  • Come Heal Yourself

    08/12/2020

    iOM Radio Network - OMTimes

    Just as we need to shower every day for our physical hygiene, we need to scrub our spiritual selves as well.Come Heal Yourself includes digging in and identifying one’s issues related and then going deep in to release those issues, followed by deep cleansing and complete healing. During this time, you will gain deep insight about your own patterns and belief system giving you Aha moments when you discover those and understand their very root. Come Heal Yourself has been developed keeping in mind, complete physical, mental emotional/spiritual healing that eventually leads to a positive, peaceful, healthy and joyful life.Listen Live at 10 PM ET Sunday on http://omtimes.com/iom

    Clean
  • Complete Estate Planning

    08/06/2020

    Nick Rosenbauer

    Do you know what will happen to your loved ones when something happens to you? If you don’t know the answer, or don’t like the answer, then this is the show for you. Listen up as we teach you about protecting your family legacy through better estate planning. The host of the Complete Estate Planning Podcast is Attorney Nick Rosenbauer of the Cincinnati area.

    Clean
  • Complete With DLC

    08/12/2020

    Complete With DLC

    The show where two bickering co-hosts get together twice a week to argue about everything video games. -------------------------------------------Hosted by Jacob Griffin and Brad Knight Follow us on various Social Medias! Jacob: Every thing is @ItsJayGriff Brad: @Kalgne on Twitter, and Brad#1261 on BattleNetFollow the Show:Twitter: https://twitter.com/CompleteWithDLCTwitch: https://www.twitch.tv/completewithdlcYoutube: https://www.youtube.com/channel/UCcmY6wtryW3PpMJPXirCuYAGoogle Play:https://goo.gl/app/playmusic?ibi=com.google.PlayMusic&isi=691797987&ius=googleplaymusic&link=https://play.google.com/music/m/Ierxqmedbc5na376wk6wfdnfcse?t%3DComplete_With_DLC

    Clean
  • Complete EM Podcast

    08/12/2020

    Complete EM LLC.

    Join George Whitney each week as he speaks with people making a difference in emergency management. They talk about what works, what doesn't, how to work efficiently, and how to get it done with maximum effect.  Special thanks to stefsax / CC BY 2.5 for our intro and outro music.

    Clean
  • "Complete Chaos" With Greg Krutzky

    08/12/2020

    Greg Krutzky

    This podcast is complete chaos...

    Clean
  • Complete The Quest

    08/12/2020

    Matt Horst

    Hey everyone! My name is Matt and I am a former Radio DJ. I am an avid lover of all things video games. Since as far back as I can remember I've been playing video games. Some of my fondest memories comes holding a controller in my hands. This channel is dedicated to talking about all the Switch games I've been playing and all the Nintendo news i'm excited about. I'm also a streamer! Make sure to check out Mixer.com/completethequest and/or Twitch.com/completethequest.

    Clean

Episode Description

Simple systems fail simply. Complex systems also fail simply, but their interconnectedness with other systems makes mitigating failures much more complex. Past a certain level of complexity, system failures are an emergent property of the system – that is, the set of system parts has a set of failure cases that the individual parts do not have by themselves. This means that it is more difficult to predict what can go wrong with a system. At some level, prediction is nearly impossible. However, you can predict many of the things that are likely to cause problems, simply by engaging in a few fairly simple thought exercises, you can greatly reduce the number of unexpected problems that your system encounters.

While it can be tempting to wait until a problem occurs to try to mitigate it, this is unwise in a production system that other people are dependent on. A system failure usually costs money at a minimum, and the problems can be far more severe than that. As a result, it's common for software services to include a Service Level Agreement or SLA, that dictates expectations about the frequency of system outages, response times, and time expected to complete work. Even if your system is engineered so that it doesn't completely fall over when a problem occurs, it can still violate an SLA and cost money. The consumers of your application probably have their own clients who have their own expectations. SLAs tend to bleed inward from clients to the services that they use and then to the services that those services use.

In contrast to SLAs, systemic problems, including both errors and latency tend to bleed outward from one service to its clients and then to the clients of that service. As a result, when you are thinking about how to find potential systemic problems, it's often best to think of these problems from two different angles. That is, you need to consider how errors and latency will bleed out as a result of a problem, while also considering how SLAs bleed in to put more stringent expectations on your system than you might expect. In effect, you are dealing with a balance between tolerance for errors and difficulty in error mitigation. Depending on how critical your system is to your clients, these expectations will vary.

You can't prevent every problem in a system, but you can usually prevent a large percentage of them by planning ahead. However, until you've encountered enough unexpected problems, it can be difficult to envision how something can go wrong, or even have a realistic thought process for thinking about what can go wrong. However, if you go through the thought exercises we've outlined here, then you have a good chance of preventing most of the problems that will plague a complicated application. While this doesn't fix everything, it can give you enough breathing room to fix the truly unusual problems that you'll occasionally encounter.

The post Preempting System Issues appeared first on Complete Developer Podcast.


Hosted on Acast. See acast.com/privacy for more information.

Discussion (0)

Be respectful and constructive in your comments