“And at our scale, humans cannot continuously monitor the status of all of our systems.” -Netflix DevOps Team
That’s why Instana came up with Stan, your virtual DevOps expert.
Why is it time for Stan?
Let's get back to basics and ask, why monitor your systems and applications? The answer is simple, as the person responsible for making sure your systems and applications are working, you need information. The hope is that monitoring tools give you information about the quality of service, help diagnose what is causing problems, then assist you by telling you what to do to fix those problems.
Unfortunately, the reality is not so pretty! Monitoring tools collect data. Data is not information. And even when configured properly, monitoring tools are not good at showing you what exactly is causing the problem, and they certainly do not give you suggestions about what to do to fix problems. Netflix describes this in their blog this way:“Many of our current insight tools are systems-oriented. They are built from the perspective of the system providing the metrics. This results in a proliferation of custom tools and views that require specialized knowledge to use and interpret.“
What this means is that with current APM and system monitoring tools, you need highly qualified experts to understand the data these tools provide. Experts are able to correlate the data and identify the root cause of a problem. But we all know, few companies have enough of these experts.
Applications today are complicated! Hundreds of servers, thousands of virtual's and containers, new releases every day or hour, constant change and churn on exactly how the application delivers its service. Instana believes that now is the time to apply the services of a highly capable virtual DevOps assistant who can help you keep things running perfectly.
This is why be built Stan! Instead of drowning in meaningless alarms, sifting through hundreds of metrics across thousands of servers, Stan directs you to exactly where the issue is and assists you with suggestions on what to do to solve the problem. In other words, Stan gets you right to the point, solve the problem!
Who, or is it What, is Stan?
Stan is a virtual robot that helps you monitor and optimize your IT systems and applications. Stan is born an expert, is always learning, never forgets, works 24x7, and never leaves your company. Stan has been gifted with years of DevOps knowledge so he understands and can help with traditional application and infrastructure environments, so called LAMP and 3-Tier Java/RDBMS architectures. But he also is uniquely capable of understanding the most cutting edge application components and architectures like the SMACK stack (Spark, Mesos, Akka, Cassanda, Kafka). In fact, Stan is loaded with unique capabilities never seen before in the world of system management, let’s go deeper into what Stan does and why he is so useful as your assistant.
What does Stan do?
Stan has a unique approach to Application Performance Management:
- Automatic Discovery and Sensoring Technology
Our agent automatically discovers all the processes and technologies Instana can monitor. For each detected component, we provide a specific sensor that knows which metrics to collect and which instrumentations are needed to derive the health of that component. All metrics are collected with one second granularity and processed in realtime.
- Dynamic Dependency Graph
Our realtime analytics platform derives a Dynamic Graph from the information the sensors are sending. This graph mirrors the runtime components and their connection interdependencies (also known as your application) and updates dynamically (in under 1 second) on any change that occurs. All metrics and measurements are applied to the graph to have a complete understanding of the current status of the monitored application.
- Health Prediction for components and systems
To understand the health of your system, we detect application and service boundaries in the graph and calculate a health status on each node of the graph. Health status is based on machine learning, mathematical algorithms and semantical knowledge about each component. When the health of any component changes, the impact on dependent components is analyzed to understand correlations and the root cause of a problem. This is used to not only understand the current application health, but also to forecast health.
- Runtime visualization and notifications
Our unique visualization based on a 3D system map is dymically updated within seconds of changes in the real application. Metric data and dashboards have 1 second granularity. Notifications occur within 3 seconds of degradation.
- Optimization suggestions
Using our knowledge system and machine learning approach, we suggest actionable optimizations for developers and operators including the root cause of each problem/optimization.
These five key technologies empower Stan to offer valuable assistance! Stan tackles the opacity of rapidly changing modern applications using our Dynamic Graph technology. Stan monitors production applications, notifies concerning QoS issues, then follows with immediate suggestions on what to fix including all the context needed to solve the problem.
Click the "Experience Instana" button to see our current demo,