We Say No to Friday Releases

By Friday we mean holidays, days off, vacation or even maternity leave. Any day prior to the period when no one will care.

Just face it, no process in the world prevents bugs. Any code contains bugs, you just don't know about them yet. What's really important is how quickly those bugs can be detected and recovered from. Resolve time means the world that's why Friday releases is not a good idea.
December 27, 2018

A Golden rule that saved many lives

Never change anything on Friday if you don't have someone to monitor and fix whatever happens. No deploying, no releases, even no preference changes – forget about it if you want to keep your weekend safe.

This rule is not questioning your professionalism, it's just about common sense. You can be 100% sure nothing can go wrong and it may not, but if it will, what's Plan B? Ask yourself, are you ready to bet your reputation against the code?

The threats of Friday releases

Here's just a short list of what can happen while you or your team are away:

  • Front-end errors in the storefront
  • Back-end errors and fails of sensitive processes (like transactions)
  • Burning support on weekends
  • Painful backups
  • Extra spendings on developer working on weekends
  • Weekend downtime
  • Business interruption and lost revenue
  • Reputation loss and damage to your brand
In addition to that, the errors in certain cases may create vulnerabilities for breaches. For two or more days your business will stay exposed to attacks and financial damage.

Again, this might happen and might not. The chances are 50/50. The Golden rule is asking, is the release worth that?

If you like to live dangerously

If there's no way to convince you to give up Friday releases, we have only one advice for you – set up a proper monitoring system. It will tell you when the store is down, or when something's wrong with resources utilization.

The one we use for AWS Cloud hosting clients monitors almost two hundred parameters of the servers 24/7/365 and alarms us for problems. Thanks to log files we know exactly what happens on every server and can quickly find the cause of the problem, fix it or roll back. The system also watches for security issues – suspicious activities or accesses. You can call us overprotective, but we have another monitoring system that watches the primary monitoring system.

Even having these precautionary measures we never deploy on Fridays. The following video is fun but it tells the story which you should remember every time you think about releasing anything on Friday.

Why choose ASAP Lab for DevOps services?

  • We are our own clients and have long ago integrated DevOps with all internal processes.
  • We have a significant experience to maintain the entire infrastructure for Enterprise businesses
  • We know how to provide development service by Agile methodology.
  • Our experience in operating complex projects includes serving leading retailers globally. We know that operation is not only about providing support but also about planning, implementing the architecture of the entire infrastructure, scaling, technology consulting, etc.
  • For years, we have been using Docker and know how to configure it correctly for projects of various sizes.
  • We are serving different platforms using PHP.
  • We have an in-house development department and a separate DevOps team.
  • The infrastructure can be either deployed from scratch or transferred (migrated) from any existing.
  • We provide services regulated by service agreement (SLA), under which we guarantee the required level of availability.