How to get cpu, memory and disk information about service fabric instances (maybe using REST apis)? [closed]

Question

I would like to have some hints to undestand the best and most convinient way to build a custom web page to monitor my service fabric cluster status showing some information about cluster nodes, for example cpu usage, free disk space, memory usage and state of single services.

Answer 1

You could use EventFlow to setup collection of Performance Counters and have that running on a stateless service on each node. From there you could collect which ever Perf Counters, and other health data you generate yourself, in addition to Service Fabric generated health data (such as Diagnostics and performance monitoring for Reliable Actors and Diagnostic functionality for Stateful Reliable Services.

Once you have setup your inputs to collect events and metrics you want to display you could create you own Custom Output that send the data to some storage or service from which you can generate your custom web view of it.

Another way would be to setup an Application Insights output and customize your dashboard to show the information you want. You could even setup export to Power BI if you want to work with that as a visualization tool and them embed your reports/Dashboards in a web page.

Or, you could, as suggested in the comments setup OMS/Log Analytics for Service Fabric.

If you're able to monitor and measure this information within your services within the cluster, you could also use Metrics and Node Capacity to gather an report this information directly within Service Fabric. You may want to do that anyway to ensure that Service Fabric is taking capacity into consideration and balancing the consumption of these resources.