As a manager, figuring out how to talk to your engineering teams about building a strong observability strategy can feel overwhelming. But don't worry! This post will help you navigate the challenges to unlock the full power of observability in your IT environment.
Drawing on insights from over 40 discussions with larger enterprises, we've put together a strategy assessment that examines three key focus areas — what we’re calling aspects — each encompassing three actionable steps.
The three key aspects to discuss with your engineering teams are:
Embrace the Human Element: This involves determining who to involve in the observability strategy, facilitating knowledge sharing among team members, and utilizing collaboration tools to enhance teamwork if necessary.
Streamline the Process: Focus on rethinking issue remediation processes, exploring innovative methods for implementing observability strategies, and ensuring alignment of all requirements and perspectives.
Expand Technology Coverage: Look beyond Kubernetes to include databases and message queues in your observability scope. Additionally, explore new approaches to effectively correlate data from various sources.
Embrace The Human Element
Catering to Various Roles
If building a strong and reliable application landscape was simple, we wouldn't see so many outages. Even the most modern engineering teams sometimes have issues. And the harder it is to pinpoint the cause, the more people need to be involved – from app owners to business stakeholders, and from devs to platform engineers.
Recognizing and addressing the distinct observability needs of each group is key. That means understanding what platform engineers, developers, app owners, and business folks require to keep things running smoothly. It's a strategic move that can make all the difference.
Here's a topic for discussion with your teams:
How can we tweak our observability tools to fit the different needs of our team members, such as platform engineers, developers, and those monitoring from the control room?
What StackState can do for you:
We’ve developed our platform in a unique way that makes it effective for team members of every knowledge level, skill set, and perspective. With the StackState observability platform, we've got you covered from start to finish – whether it's end-to-end monitoring, troubleshooting Kubernetes issues, or diving deep into application insights. Explore our end-to-end observability solution for more details.
Manage the knowledge
From our talks with these 40+ companies, one thing became clear: valuable knowledge tends to stay confined to the experts. It's crucial to capture and share that knowledge across the organization. By integrating it into your observability platform, discussions can shift from "how" to "why." This will drive the broader understanding of all engineers to the next level.
Our insight? Make observability policies a part of your toolkit, just like we did with CI/CD and infrastructure as code.
Here's a topic for discussion with your teams:
How can we ramp up our knowledge-sharing game and make sure everyone's expertise benefits the whole team?
What StackState can do for you:
The StackState observability platform comes with over 35+ pre-packaged monitors right out of the box. These cover your initial set of policies to boost your reliability. Plus, we've made it easy to expand our monitors if your engineers require more extensive coverage.
Foster Collaboration
When there's a problem or, worse, a full-blown outage, that's when the learning kicks in. Teams can band together to fix things fast and then dive into a solid post-mortem to learn from the experience. It's all about reflecting on what happened and taking steps to prevent it from happening again.
Bridging the gaps between teams is a must in any observability strategy. It ensures that observability becomes a team effort, boosting efficiency and putting an end to the blame game.
Here's a topic for discussion with your teams:
What steps can we take to improve teamwork across different departments and make sure everyone contributes to our observability efforts?
What StackState can do for you:
Having all your data centralized in one place is a solid foundation, and that's exactly what StackState provides at its core. But we go even further by giving you the power to replay any changes that occurred in your cluster, step by step and over time, so you get the bigger picture. Check out how we map component dependencies in your cluster.
Streamline The Process
Accelerate from detection to resolution
Learning to navigate the quickest path from identifying issues to resolving them is crucial. The SRE Handbook outlines a clear process for this, yet there's not much documentation on how to tackle it. Having all the data in one spot, with the right guidance on where to focus, is fundamental. Ask yourself: How well does your current solution guide the remediation process?
Here's a topic for discussion with your teams:
What strategies could we adopt to speed up our issue detection and resolution process?
What StackState can do for you:
StackState superpowers your team's ability to pinpoint the root cause of slowdowns or service interruptions. And with pre-packaged monitors and remediation guides , your teams are quickly guided to the best solution.
Shorten the time to initial value
Choose observability tools that teams can easily adopt and enjoy using, bringing instant benefits to your operations. This is one of the biggest takeaways we've gathered from numerous company discussions. While open-source solutions may suffice and even excel in certain aspects, it's important to consider where you want to invest your team's time and efforts.
Opting for a stack of individual solutions means investing time in four key areas:
Selecting the right tools. This may be straightforward for some aspects like dashboarding with Grafana, but less clear for metrics storage options such as Thalos, Prometheus, or VictoriaMetrics (StackState utilizes the latter under the hood).
Configuring the tools and assisting teams in integrating their business applications into the observability stack.
Prioritizing training and education. Mastering a suite of diverse tools takes time and effort to empower your engineering teams.
Considering innovation. Decide whether you aim to innovate in observability or treat it as a standard commodity. This is a crucial question to address.
Here's a topic for discussion with your teams:
How can we choose observability tools that are easy for our teams to adopt, use, and immediately benefit from?
What StackState can do for you:
StackState connects the dots between metrics, events, logs, traces, and change information. We've also refined a super user-friendly navigation system that many companies find eye-opening. Take a look at how StackState can correlate data.
Tighten integration between different perspectives
Imagine if you could synchronize different views from the control room to the developers. This would help ensure a unified approach to problem-solving and bring order to chaos. Well, in speaking with organizations in the market, we've identified three distinct approaches to observability and monitoring.
End-to-end observability. Network operation centers and master control rooms need to see the big picture. They’re seeking to find out how their business applications are connected, and what issues might cause a disruption in the business process.
In-depth application insights. Your development teams often request advanced application monitoring to help them optimize performance and understand slow traces, among other things!
The tech stack below. This is typically of interest to infrastructure or cloud platform teams.
The interesting thing about these three different approaches is that they all draw from the same data. Despite using the same technical resources, they offer varied answers and perspectives. It’s super important to understand this concept thoroughly and base your observability strategy on it.
Here's a topic for discussion with your teams:
How can we better integrate the various viewpoints within our team, from the control room to the developers, to solve problems more effectively?
What StackState can do for you:
Our services are aligned with the key players in this field. Take a look at our 3 solution pages on how we’ve improved Kubernetes troubleshooting, end-to-end observability, and application performance monitoring.
Expand The Technology Coverage
Beyond Kubernetes
Realize that true observability extends beyond just your Kubernetes setup to encompass databases, endpoints, message queues, and beyond. Life would be simple if cluster borders marked the extent of your responsibility. However, there's much more that needs to come under the supervision of your observability solution.
Here's a topic for discussion with your teams:
Given that our observability covers databases, endpoints, message queues, and more — not just our Kubernetes setup — what's the most effective approach to monitoring everything?
What StackState can do for you:
With a shared data fabric in the background, StackState can ingest observability data from any source. What’s more, the data exchange has been streamlined through the adoption of standards like OpenTelemetry and eBPF.
Beyond health
If basic health signals were all we needed, we could get started right away. However, there are other areas worth considering. Compliance and security are probably top priorities. Plus, linking your CI/CD tools or CMDB to your observability stack can provide valuable insights into the impact of changes.
And shouldn't we also include business metrics or user experience metrics? Diversifying the types of indicators monitored can strengthen your observability strategy and improve team response and performance.
Here's a topic for discussion with your teams:
How can we include a wider range of health indicators, like security alerts, in our observability strategy to improve our team's response?
What StackState can do for you:
Integrating with other tools is seamless. StackState offers integrations with security tools like PrismaCloud or Falco, as well as CI/CD pipelines like Tekton. And we keep you informed through a variety of notifications and alerts to provide a more comprehensive view of the overall picture.
Beyond distinct data sets
Move beyond isolated, siloed data. Focus on connecting metrics, events, and traces. You'll find that simply having data is just the beginning: without context from correlated data, making sense of it all becomes tricky, and your engineers may find themselves spending more time navigating through data than reaching conclusions.
Here's a topic for discussion with your teams:
How can we break down data silos and merge our metrics, events, and traces for a more comprehensive view of our systems?
What StackState can do for you:
Connecting data is truly StackState’s secret sauce. The reason we excel in other areas outlined above is that the data is connected through topology and is super easy to navigate due to the fact that all telemetry data is correlated. Learn more about StackState's full-stack observability platform and what sets us apart.