Phoenix Project: Sometimes you have to look back to look forward
It has been eight years since The Phoenix Project was published and a lot has changed since then! I started to think about what we’ve learned in that time.
It starts with the theory of constraints. I still see it all the time. Organizations take actions which are merely temporary, putting out fires but not solving for the underlying causes of those fires. The book talked about specific ways organizations could focus on constantly identifying and removing these bottlenecks and persistent issues via the Three Ways.
The path to success via The Three Ways
Way 1 – Move to smaller intervals of work, remove constraints, identify ways to improve delivery constantly.
Way 2 – Ensure fast detection of quality issues, get early feedback, prevent recurrent problems.
Way 3 – Use the efficiencies from #1 and the safety from #2 to accelerate innovation, and a culture of thoughtful risk taking and experimentation.
If you want a quick refresh of the Ways, check out this video from Gene Kim. They still hold up, even with all of the changes IT teams have faced in the last decade. Much of the change has been very good for IT even as new issues have arisen. Continuous Delivery is now a given for many organizations, velocity has accelerated, and change is a constant. That can be a problem.
Enterprises have invested in solutions to help them better manage the associated risks, and are now monitoring everything, leading to tens of thousands of performance alerts per day. For many, lack of visibility is no longer an issue. Their NEW challenges are alert storms, too much data without enough intelligence to act quickly, lack of consistency, siloed teams and approaches.
What has not changed:
- Ongoing reliance on tribal knowledge, which creates bottlenecks because of a never-ending dependence on SME’s
- Persistent issues with performance and availability
- Ballooning operational costs combined with organizational waste as too many resources are involved to triage issues
Enter DevOps…and things got even more complicated.
While there is no question that the creation of a role intended to bring developers and operations closer has been a step forward, DevOps did not solve all our problems.
In large organizations, the distributed nature of DevOps can lead to great difficulty operating the multiple services required to run an enterprise. Decentralized groups optimize monitoring at the team level, often duplicating efforts across teams, proliferating tools and silos, and still getting pulled into bridge calls and war rooms.
Meanwhile, centralized IT Ops struggles to achieve standardization across the board, a dream that for most can’t come true. We are left with a big gap to fill and this question to answer: how do we enable DevOps to run lean, fast, and localized while allowing IT Ops to have a centralized and efficient approach across the entire landscape?
The future is here
Enter BigPanda and AIOps. BigPanda takes a tools agnostic approach that enables distributed teams to leverage the solution most effective for their needs, while providing IT Ops a centralized and unified manner to standardize incident management processes.
By cross-correlating the wealth of data across all monitoring, change, and topology sources BigPanda allows IT Ops to reduce the signal to noise ratio, enrich incidents with important metadata to remove constraints of tribal knowledge, prioritize response, and leverage a platform for automation that ultimately helps improve performance as well as reducing operational workload and cost.
How does this apply to the Phoenix Project and the Three Ways?
Way 1 – Remove constraints from the IT Ops team being able to solve problems quickly.
Way 2 – Identify when an issue is impactful quickly, leverage analytics to pinpoint recurrent issues and solve them permanently with thoughtful change or automation.
Way 3 – Improved ability to solve problems quickly, using fewer resources, frees up teams that should be focused on innovation to spend more time delivering against their mission. It ensures that as innovation is delivered and risks are taken, IT Ops can easily identify, prioritize, and resolve any incidents that cause a negative impact.
The BigPanda approach to AIOps gives you the best of both worlds:
An agile DevOps approach that empowers individual teams and their services alongside a consistent and silo-free, centralized approach to IT Ops that ensures efficiency in incident management. The combination improves your ability to manage performance and availability, and reduce operational workloads so teams have time for innovation work. The outcomes are improvements in MTTR, time saved from firefighting reallocated for innovation, empowering IT Ops to be a leader in leveraging automation, and ultimately revenue protected, and dollars saved.
Don’t take my word for it though, watch any of the on-demand sessions from our recent conference and hear it directly from the IT Ops community.