Achieving the right, healthy balance in life can be similar to achieving the right balance in your IT organization. There are multiple variables that make both a challenge. In life, it’s about the right mix of nutrition, fitness and healthy choices, not to mention family, finances and personal goals. In IT, the mix starts with people, process, and tools, but also includes costs, time and business alignment.
BMI’ is one metric used to track your adoption to a good balance in life. Likewise, assessing your adoption of DevOps practices can be a good metric to tracking the proper balance in IT. In light of the challenges that many organizations are facing to achieve the right balance of DevOps, IDC recently released a survey that tracks progress of and barriers to implementing DevOps. They found that IT teams advanced with DevOps adoption in different dimensions at different times: people, culture, technology, business and process. While the people, technology, and process elements evolved along a common trajectory, the culture and business sides prove to be more complex.
When implementing DevOps practices and guiding teams to automate incident management, there are five major areas that IT leaders should consider to make sure the right balance is achieved between business needs and cultural development:
1. Assessing your delivery lifecycle today, and determining what you want it to be.
While initial focus of DevOps is typically focused around the automation of software builds to speed delivery, integration testing, and code deployment, the next phase of DevOps adoption will focus on the automation of monitoring, analytics, and remediation. The increasing sophistication of IT Operations Analytics (ITOA) tools can help drive organizations’ ability to quickly remediate problems during new code rollouts, correlating application performance with change-related data and the performance of the underlying infrastructure. As you progress from development to QA to production, you can allow room for experimentation by converging visibility across Dev and Ops tools and clustering data that may indicate a potential issue upon code deployment. This involves understanding the ecosystem by identifying and acting on relevant data across the application service stack, which enables DevOps teams to catch service affecting issues at the onset.
2. Building the capability to detect failures earlier.
DevOps teams can be overwhelmed with all the alerts that are generated by the various monitoring tools. Many are familiar with “alert fatigue,” e.g. getting continually notified in the middle of the night when less-than-critical alerts are sent. But worse is not knowing about service-affecting situations until it’s too late. Applying analytics to the large variety of event and alert streams can be a Goldilocks dilemma: not too many false negatives, not too many false positives, but just the right amount real problem notification. This can only be achieved through tools that “learn” over time.
Real-time monitoring that uses machine learning to pool relevant data from across the IT application service stack and adapts to diverse infrastructures (software-defined, hybrid cloud, etc.) is key to faster incident detection.
3. Managing the volume of alerts and distribute notification.
Machine learning enables clustering, correlating and contextualizing related alerts (homing in on relationships between event data) and converting it into a single view or “situation.” Through this consolidation of related events, teams will be better equipped to home in on and diagnose issues at scale.
4. Improving collaboration to diagnose and remediate faster.
Five to 10 years ago enterprises tried social IT to solve problems collaboratively. Unfortunately, It didn’t work. Project participants still track their side of the story in separate systems, using emails or conference calls. Now enter DevOps, and the need for a collaborative support system generated on-demand to solve problems. In 2015, look for services that create this single support space for not only IT resolution but for sharing of tools, diagnostics and knowledge.
To be effective, teams need a “shared workbench” to troubleshoot issues by designating tasks and leveraging the event history of respective service-affecting incidents. The “single pane of glass” counters the lack of visibility on the impact of transient outages in other silos as well as duplication of efforts during the recovery process.
5. Facilitate continual experimentation and learning.
Traditionally, Dev teams would deliver code to QA teams, who would then test it, while Ops teams would (with a different toolset, and most probably, manually) deploy the app. But with the introduction of the “Cloud Aware Application” model, instrumentation and self-testing are embedded features and so, there’s a native tolerance for a high rate of failure. This, in combination with A/B testing on a small percentage of the user base, allows for a quick roll-back if a deployment is not successful.
This approach also requires new monitoring tools as well -- those that help assess the relevance and accuracy of log messages in Dev/Qa prior to production roll-out, and reveals the dynamics of the interchange between the application and the underlying infrastructure.
Where do you sit along this spectrum today? If you’re seeking to deploy application-services faster, all while automating many of the operational process of support them (including incident management), we’d love to know how these factors support your efforts to advance the culture of your DevOps team. The right balance of people, processes and tools can bring your Dev and Ops teams together, avoiding the “The Biggest Loser,” and getting to the “The Amazing Race.”
Rob Markovich is chief marketing officer at Moogsoft.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access