Have you been tasked with getting monitoring into your application and struggling to figure out where to start?

Let me share some github repos where IBM has contributed open source data collectors that you can use to instrument your application to collect resource and performance monitoring data:

Java: https://github.com/RuntimeTools/javametrics
Node.js: https://github.com/RuntimeTools/appmetrics
Swift: https://github.com/RuntimeTools/swiftmetrics

Lets look at the Node.js data collector! The overhead on your application is shown to be minimal, adding less than 0.5% to the CPU usage of your application. Additional memory requires about 20MB to gather information about your system and the application. It is easy to get started and to collect information from your application while developing it, by using the following two procedures.

  1. You can use your browser to point to your localhost port 3001 to view a dashboard which shows you performance metrics for you to tune your application’s resource usage.

    Note: If Eclipse is your preferred development environment, you can plug in a health center client and additionally get profiling data about your application.

  2. This is where IBM’s Cloud App Management (ibm.biz/monitor) tool comes in to help.

    IBM Cloud App Management collects data from your open source data collectors and shows the same metrics you used at development time, so operations and development have the same familiarity with how metrics should look. Additionally, Cloud App Management maps the metrics into a standard set of four key golden signals (latency, errors, traffic, and saturation) which helps operations with a common view of your service vs other services, which may be written in a different language. This is a common practice amongst those who adopt site reliabilitiy engineering (SRE). They adopt these key signals that can be the base of your service level indicator.

    We encourage you to start with these signals, but then you should define additional SLIs to suit your specific business objectives. Since the data collector is open source, you can branch it to add custom metrics to customize your service level indicators (SLIs).

Writer: Mike Mallo
Editor: Carmel Burgess

Join The Discussion

Your email address will not be published. Required fields are marked *