Demystifying Service Observability!

😩 Pain point

Modern systems can emit millions of metrics & modern monitoring tools can collect them all. Faced with such an abundance of data, it can be difficult to know where to start looking for an effective service monitoring.

😇 Solution

Think service observability through it’s stakeholders. This leads us to think in two different perspectives -

💁‍♂️ External customer’s perspective.

This perspective is Request scoped & here one should consider using the RED strategy! It’s an externally-visible view of the behavior of the workload as serviced by the resources.

“RED” acronym stands for Rate, Errors, and Duration.

• Request Rate – Rate at which requests are coming
• Request Error- Number of requests that cause an incorrect, incomplete or unexpected result
• Request Duration- Amount of time the request took

💁‍♂️ Internal Operator’s perspective

This perspective is Resource scoped & here one should consider using the USE strategy! The goal is to understand how these resources behave in the presence of the load.

“USE” acronym stands for utilization, saturation, and errors. USE directs attention to the resources handling the workload. For every resource, check-

• Resource Utilisation – Average time the resource was busy servicing work
• Resource Saturation – Degree to which the resource has extra work which it can’t service, often queued
• Resource Errors – Count of error events for the resource