Someone asked me to explain Full Stack Observability this week and I figured I'd share my answer here. Traditional monitoring was performed in silos. Server folks monitored servers, application folks monitored servers and network folks monitoring, well, the network. Anyone wanting to understand an end-user's experience had to look at each of these silos independently with no real correlation unless you could do something in Excel manually. Introduce modern technologies like Cloud, Containers, virtualization, etc. and this model breaks down quickly. Complicating the situation is the number of tools required to run an IT operation - syslog, server monitoring, workstation monitoring, network monitoring, cloud monitoring, container monitoring, service availability, etc. Also, traditional polling-based monitoring is being augmented (if not replaced) by device telemetry.
Full Stack Observability focuses on providing a view across all layers of the OSI model based on persona's. For instance, an application owner can see any and all incidence and performance information across the entire stack. This view will differ from a service owner who may have multiple applications and need a broader view. For some, this many sound like a manager of managers, which I would argue was the 'infant' version of Full Stack Observability. FSO leverages the power of AI, ML and modern correlation tools to bring a business-focused view of the environment.
The other interesting note: Full Stack Observability is on the evolution path to AIOps. AI will be leveraged allowing the environment to predict and prevent issues and to self-heal when issues arise. For more information, here are a few good links: https://www.appdynamics.com/topics/what-is-full-stack-observability https://www.dynatrace.com/monitoring/platform/observability/?utm_source=google&utm_medium=cpc&utm_term=observability&utm_campaign=us-trending-cloudops&utm_content=none&gclid=EAIaIQobChMIpMOrgNyG8wIVtHxvBB0tNQUiEAAYAiAAEgI4-fD_BwE&gclsrc=aw.ds
Comentários