Walking around at one of the companies I advise and help with their DevOps challenges I heard the terms above used in a ‘mix and match’ manner resulting in fuzzy communication. Aiming to reduce confusion and improve understanding between people, I thought it might help to write out the distinction between the three.
Monitoring – Is our platform operational y/n?
For instance, is our software platform handling requests at this very moment. This is monitoring in the most basic sense and the answer to this ‘yes or no’ question says whether we are still servicing our customers or not.
Data analysis – How many users performed function X yesterday?
This is a data analysis question. Depending on the complexities of the platform and the analytical questions, this information is based on accumulated and aggregated data. This data can be centralized to one location in the form of a datalake for visualizing the data in dashboards and or e.g. for generate customer reports.
Logging – Does our application produce any errors?
This is conventional application logging. It usually ends up in a system log somewhere. It is used to analyze weird behavior in the application’s functionality. Note that this can also be partially used for data analyses. However it’s main purpose is for debugging issues.
These information flows are and should be handled in distinct processes and systems and are most likely valuable for different departments and/or audiences.
In addition, as pointed out by my favorite data analytics guru Thierry Fakkert, Monitoring is about ‘now’ and Analytics is about ‘history’.
Good luck and keep your data sane! 😉
-Gerard.
