Can IT admins benefit from a DevOps environment?
If DevOps is implemented poorly, it increases the pressure on admins to deal with uncontrolled code rollouts in the operational environment, which can drain hardware resources and increase troubleshooting time.
However, a good DevOps environment with all the right checks and balances in place reduces the number of mind-numbing tasks that admins face, and it provides more time for interesting, value-add tasks that help the business.
DevOps also gives admins oversight of developers. Powerful DevOps tools place all the correct filters on the coding and application development process to abstract the raw developer environment from data center operations.
Are there benefits to a DevOps environment?
DevOps is based on emphasizing better collaboration and communication between developers and operations professionals. To help improve collaboration, many organizations turn to advanced automation; it introduces new, more streamlined ways of moving data center-supported software and applications. A DevOps orchestration tool can support the intelligent implementation of patches and upgrades on existing commercial off-the-shelf software and open source software. This helps keep software up to date without all the trouble of using manual scripts and long prechange testing cycles.
Automation frees up data center admins to focus on things that are far more important, such as how to adopt hybrid cloud; how microservices and containers may affect the overall data center architecture; and how to implement and manage the move to a more dynamic cost- and value-based workload management system.
These are the things that directly affect the business and are worthy of investment rather than firefighting issues introduced through manual actions, such as incorrectly flagged security events or mismanaged data center resources.
Set up testing
DevOps tools help produce well-written code, and developers carry out both pseudo and real code testing to uncover any issues that can appear in the live environment. This is done through either creating a discrete scaled environment for testing in a safe area of the application or using parallelism in a live environment and redirecting some of the live traffic to run through the test code.
To help build these test environments, admins must confirm with developers how much storage, memory and processing are necessary and set up any partitions. Depending on the type of testing, developers can also require virtualized resources or GPUs if they must run a large amount of code or extensive data sets.
Create a feedback loop
A DevOps environment integrates any feedback received at either a technical level via metrics or directly from users via the help desk. This setup ensures that developers get rapid, effective visibility of any code-based issues but also the priority level of the issue.
This eliminates the need for data center admins to knock down the developers’ door and try to impress upon on them the urgency and severity of any production issues that appear in the data center; with a DevOps environment and systems in place, there is equal visibility for both departments across the business.
Good DevOps systems also introduce full workload orchestration. Admins and developers can set the software to automatically implement a system reset or rollback should a problematic piece of code appear.
Another example of workload orchestration is elastic — and automatic — resource scaling if any of the current setups constrains new application code. Admins can also help developers properly delegate workloads so applications can run on the best-suited and most cost-effective infrastructure.
If it turns out that a workload is on an expensive public cloud but admins find a cheaper pricing structure and appropriate resources, then they can move the workload to the better option as required.
This type of workload automation goes beyond what most data center admins are used to semistatic, proprietary workload managers that specialize in one area. DevOps offerings from Electric Cloud, HashiCorp and Stonebranch give the user a more holistic view of all integrated workloads and support continuous development.