Bad advice for your startup

The third part of the "History of a startup" is delayed due to suddenly happened holidays (who did not read - this is the beginning), here is a set of harmful tips for you. With "History ..." they are not connected in any way, just observing various projects in which I had the opportunity to participate, or my colleagues.

Council number 1. Take computing power with a margin. After all, your application is a breakthrough in the industry, and the number of visitors will roll off two hours after release on the Google Market / App Store. Take the server under the advertising site and the server under the Ansible controller, and also follow the recommendations of the software manufacturers that the deployment is less than three servers - this is not a production level.

Council number 2. Use as much SaaS as possible. It is desirable - with the absence of a simple mechanism for moving from this SaaS to your own hosting. Ideally, the solution should be paid for with some introductory period, after which most of the functionality is disabled. After all, the application is “already-almost-about-ready”, placement in the Google Market / App Store takes a couple of hours, and then you will immediately earn so much money that it will be enough to pay all the accounts, and the beer will remain.

Council number 3. Take advantage of cloud computing power. After all, when you have an application in the cloud - you do not need an architect or DevOps, in the cloud the application will scale itself, which will reduce your operating costs.

Council number 4. When it turns out that everything is not so simple with the clouds - give programmers the task of designing the application architecture taking into account scaling, and automate the deployment. Do not listen to their timid hints that it would be nice to take at least some admin to the team - everything is in your cloud, there the servers themselves know when to start, when to stop, and where your application data is stored. And if they do not know, there is Ansible, for the use of which you don’t need to know at all what "system administration" is. Know the YAML configs slap.

Council number 5. Make non-switchable code exception monitoring. After all, it will always be very important for you to know how many push messages were delivered during the period "on this day, hour, minute and second two years ago." Add the data to the cloud-based SaaS service with its API - after all, storing such a volume in ordinary files is the last century and expensive.

Council number 6. Never implement monitoring at the operating system level. After all, if the application stops working, you will see it (it is running on your phone / browser tab 24 hours a day), and you can see the graphics of CPU usage and memory usage in the web-based management interface of your cloud.

Council number 7. When someone from the team gently hints at the uselessness of 60% of the movements already made in the project, and that "well, maybe it, let's first settle down on the fatter virtual machine" - wave it off, or even stop working with the person, because he is incompetent. Remember, without High Availability, Big Data and Scalability, no startup survives.

Well, share your bad advice in the comments, or something ...