Today there are two major categories of cloud performance tools in the market: active and passive.
Passive cloud management tools
Most cloud computing management tools are passive. Passive tools provide you with information about issues within the cluster of cloud systems, and will perhaps even recommend a course of action, but they do not actually take corrective action.
The idea behind passive tools is that the corrective actions are so complex it does not make sense to automate systems corrections. This is the right approach for many enterprises. Their systems often provide information that needs human interaction because the information is event-based and not operations-based, generally speaking. The passive approach that engages humans in the processes feels the most comfortable to cloudops teams.
Active cloud management tools
Active cloud management tools seem to be where the market is going, considering that we want to automate ourselves away from the complexity of cloud management. We are quickly approaching a tipping point. A modern cloud-based deployment may have as many as three public clouds, hundreds of different databases, dozens of security models, and 50 or more storage types, with more complexity coming soon. We have run out of the resources required to maintain these systems; there is no path forward without some automation and active tools.
You can divide active tools into the following categories:
Analytics-oriented active tools are similar to analytics-oriented passive tools, but they can take steps to carry out corrections or temporary fixes. An example: predictive analytics and logged data indicate a high number of network errors. It has become apparent that a core router will go down soon. Humans need to be notified of the issue so that a new router can be swapped in, but the tool could also route data around the defective router before humans even know the issue exists.
Event-oriented active tools can capture simple events just as event-oriented passive tools can, but active tools can take automated corrective action, typically simple issues that have obvious but very effective fixes. This is the majority of systems issues that cloudops engineers face on a daily basis.
Self-correcting or self-healing tools. This interchangeable term is now tossed around in the marketplace, and it is core to both analytics-oriented and event-oriented approaches. Indeed, it is typically the subsystem that carries out the automated repair. It is reasonable to assume this category of tools will rise in importance, considering the amount of work it takes to figure out how to automate corrections. The diagnostics tools could populate one side of the market, with corrections tools on the other side. Correcting tools may offer more viable and scalable options because these tools need to layer into some pretty specific cloud-native technology.
Expect that active cloud management tools will be the future. Are you ready?