There are many options when it comes to container orchestration platforms and services. Figuring out which one is best for a particular organization’s needs and applications can be a challenge. Which platform for containerization you choose can significantly influence your business success, so the selection process should be carefully considered ...
Although DBAs fortunately have the rare ability to bridge the gap between development and operations, they have been detrimentally overlooked in many companies that deploy DevOps practices. A DBA's ability to interrogate code and construct a resilient, well–performing database environment uniquely defines the capabilities needed for DevOps.
Whether through formal methods such as classroom or virtual training, job shadowing, and mentoring; or through informal methods such as team discussions or presentations, teaching needs to be a frequent element of team integration. It is a given that IT and business teams have difficulty understanding each other without a common taxonomy. Even teams within IT often fail to understand each other.
A developer discussing encapsulation or inheritance may totally perplex a DBA unfamiliar with object-oriented programming terminology. Never mind if you start talking about Agile, which is very new to many IT professionals. Likewise, a DBA ranting about developers "thrashing" the buffer cache is likely to see the "deer in the headlights" stare.
While investigating a performance issue specific to a screen, a developer shared with a DBA that the drop-down window would display ten data elements from which the application user could select. As they looked at the code and then tested the code in a nonprod environment, they learned that the result set was millions of records. The million records would move from the database to the middle tier, and then the needed rows would be pushed to the client application screen. When asking why millions of rows were being returned, the developer said that was a standard practice. After looking into other queries, the DBA soon found herself ranting to several development managers about the developers thrashing the buffer cache and the performance impact. After realizing that these managers did not understand DBA "technical" jargon, she determined that there was a better way to communicate the message. She scheduled a meeting a few days later, in which she put together a presentation deck outlining basic buffer cache concepts with visuals (see Figure 1 ) that demonstrated how large result sets can negatively impact not only the query requesting the data but also every aspect of the database performance.
Figure 1. Buffer cache thrashing
After the DBA spent an hour walking the developers through the presentation and answering questions, these developers understood the impact of less-selective queries. As days and weeks passed, and often when the DBA was visiting the developer realm, developers would jokingly remind each other to not thrash that buffer cache unless they wanted the DBA to get after them. Although the training was succinct and simplified, it closed the language gap, resulting in improved query selection criteria, smaller result sets, and less buffer cache "thrashing."
The point is that even people in the same industry do not necessarily speak the same language. DevOps introduces another language gap that requires purposeful definition to keep all members of the team aligned. This book presumes that readers are technically savvy and already familiar with DevOps and the core terminology, but it may not be true as they begin working with DBAs. Accelerating DBA engagement requires DBAs to understand the DevOps principles and foundational constructs.
Experienced DevOps team members need to educate DBAs on processes, continuous integration and delivery, and the implemented tool set. Demonstrating how code is built, tested, integrated, and released helps DBAs determine where best to interject changes supporting the code cycle. DBAs also need early notification when system changes are necessary, allowing time for the reconfiguration to be completed, tested, security approved, and automated for pipeline consumption.
Differentiating which DBA inputs to put forth for absorption into existing agile and DevOps processes demands collaborative effort between existing team members and newly assigned DBAs. Cohesive integration to advance the undertaking of capturing additional value at decreased costs lengthens the backbone — the code generation process definitions from start to finish — of the movement, triggering existing processes to be rehashed, or repurposed, and then reacclimated within the SDLC cycle.
Together, DBAs and DevOps team members make old things new again as processes throughout the development, testing, release, and operations support pathway are refined to incorporate DBA tools, change methods, and metrics. The critical goal is to not disrupt the code delivery schedule while reaffirming the automation and process sequence preciseness. Sanctioning a parallel environment that initially mirrors the primary build-to-release architecture onto which the DBA components get added enables a side-by-side comparison to ensure that updated processes work correctly. Of course, automation oversees the execution, examination, and effects reporting.
Quick to Value, Delight the Customer
Excitement for DevOps, besides the "it's the cool thing now" factor, stems from years of frustration as IT professionals have been viewed as moneywasting, unresponsive, slow to deliver, and second-rate business citizens. One of my pet peeves has been the "IT alignment to the business" language. Viewing IT as an "outside" entity having to blend in plans to support or conform to the rest of the business accounts for much disillusion and poor esprit de corps.
When agile development (and DevOps in close pursuit) exploded in popularity, IT folks finally envisioned a promising future in which product delivery proficiencies incessantly eliminate time, process, approval, and implementation waste, and then rocket delivery to the customer. One Lean principle is establish pull. Customers establish pull inherently when reporting problems or requesting new product functionality. IT's capability to deliver has never been this radically empowered, in which demand (pull) can be satisfied within a customer's time expectations.
As consolidated teams, call them agile or DevOps, build new or decouple established services from monolithic applications, change footprints become much smaller (think microservices), making it possible to deploy code quickly with minimal risk. With speed united with smaller code chunks, a failed release becomes no more than a temporary blip on the radar.
This blog is an excerpt from Mike Cuppet's book: DevOps, DBAs, and DBaaS