Minimal effort to check if something is broken
We had an unused tv at our disposal a few months back and were looking for ways to use it in a productive manner (no Netflix!). Someone came up with the idea to show all kinds of dashboards, for logging, builds, deploys, metrics, you name it. But casting from another device or plugging someone’s laptop to the tv wasn’t what we were looking for.
A small but great piece of technology we found was the Intel© Compute Stick, a device that enables you to bring a small-scale Windows computer to any screen that offers an HDMI-port. All you have to do is plug it into a monitor (or television in our case) go through Windows’ first startup sequence, plug in a keyboard/trackpad and you’re good to go!
Intel® Compute Stick is a device the size of a pack of gum that turns any HDMI* display into a fully functional computer: same operating system, same high quality graphics, and same wireless connectivity. All this in a PC on a stick that measures 4.5 inches from end to end, and is ready to compute right out of the box.Intel©
So what are we showing and why?
In our case, we separated the dashboards in groups based on microservices that exist in the same ecosystem or context as we call it.
That means that for every build we can check IF it failed and WHEN it failed, and who was responsible (pointing fingers is fun when it’s not you, right?).
This way we can minimise the time needed for someone to stumble upon a problem and get right to fixing it.
In conclusion, we are able to follow up on all the DevOps pipelines in one quick view.
Another great tool we use and show on the big screen is Kibana.
ToThePoint: Kibana lets you visualise your data that has been collected by Elasticsearch, for example all the logs our services spit out are kept there and Kibana produces graphs and tables to make something out of the cluster of content stored in an Elastic Stack.
A third, more DevOps oriented dashboard tool we utilise is a Grafana dashboard. Grafana shows all sorts of metrics pulled from our very own Kubernetes Cluster. Memory, CPU and network usage are all shown live so we can make sure that our cluster runs smoothly at all times.
Bringing the pain forward
All of this adds up to a smooth collaboration between development and operations, making life easier for everyone. But of course most important of all, it looks cool.