How automatable are garbage workers

A systematic garbage collection against too much waste in the cloud

Cloud computing is spreading at double-digit market growth rates. In the beginning it was more start-ups or internet companies that managed without their own data center, now some large companies are pursuing a cloud-only strategy. Initially, the focus was perhaps on cost advantages, but the appetite often comes from eating. More and more cloud services are used, especially for new applications. Do not lose sight of the costs. In the cloud, deep traps and opportunities are often closely related. We address the typical stumbling blocks when it comes to cloud costs and show you how you can avoid them.

The easiest way to save costs in the cloud is simply not to use cloud services. This primarily relates to environments for testing and development, which are often only used during typical office hours between 8:00 a.m. and 6:00 p.m. If these environments are shut down outside of these times alone, the monthly bill is reduced by around 30 percent.

Just as important as the automated creation of the cloud environment is the shutdown or scaling when resources are no longer required so much, such as in a test environment.

Another problem is over-provisioning. It often occurs with applications that have been migrated using the lift and shift approach. Since their actual load profile is not known or not monitored, these environments are usually dimensioned beyond the maximum requirement. Here it is worthwhile to measure precisely and possibly to choose the next smaller instance.

Over-provisioning can also mean that the booked performance class is no longer needed. This often occurs with applications that are at the end of their life cycle, as they are used less and less and ultimately only sporadically. Here, too, a performance class that is adapted to the actual use is worthwhile. Often the cloud provider also offers discounted spot prices if you can live with less availability, e.g. B. in the case of sporadic or temporary use. In addition, the cloud providers are further developing their instance classes so that it is often interesting in terms of price and performance to switch to the next smaller, but more powerful, modern instance. If the application allows it, a change from the x86 Intel processor platform to an ARM computer can be worthwhile, as these are often much cheaper to operate.

Another possibility to save operating costs is to use the managed services of the cloud provider as an alternative to the self-managed services. This allows you to use your valuable human resources for more economically important services or to optimize the cloud infrastructure.

Operations and development often have to work closely together to find the right mix of services at the moment. Comprehensive cooperation and responsibility are already a prerequisite for managing both the infrastructure and the software code automatically via the DevSecOps pipeline

Another important prerequisite for being able to optimally monitor and analyze costs is the exact allocation of the costs to the causer and the possibility of comparison with previous consumption in order to be able to estimate the need for optimization. This is the only way you can transparently understand where and how great the optimization potential is and avoid unnecessary "waste".

From time to time it is also worthwhile to use external help and know-how for systematic garbage collection in the cloud. With their multi-cloud know-how from development and operation, Materna consultants can get even more out of your existing cloud services and, for example, give you a good overview and a basis for decision-making with the cloud readiness check. We help you to continue to successfully circumnavigate the cloud cost cliffs and to be successful on the way.

To the cloud readiness check.