This week's episode of the Kubelist Podcast was a conversation with Rick Spencer from InfluxData. Influx has been running K8s in production for a while now and has the battle scars to prove it. Rick's honesty about what was challenging was a refreshing and good look.
Influx has been a go-to for some time now when in need of an easy-to-setup and relatively robust time-series database that runs well locally in a container. While you can run it in Kubernetes, it’s possible to run it outside of Kubernetes also. Don’t worry, we include a few other links in this issue to show some interesting use cases for running a time series database, if you need inspiration. ⏱️
Let’s talk about observability and monitoring, because that’s a common time-series use case. I liked this quote from the beginning of this post: “What often prevents them from implementing an effective monitoring solution is the plethora of choices they face.” You can call this analysis paralysis or whatever you’d like, but we all know that it’s better to monitor than not. If you don’t have a monitoring solution in place, here’s a quick path to a good system that’s not going to take your SRE team weeks or months to configure and install. 👀
Telegraf is the collector/agent part of the Influx stack. This agent is open source and includes support for a lot of data sources via the plugin system. Looking at this list of plugins might inspire you to build a quick application that depends on external data. Run a copy of InfluxDB (or just use the cloud version since it's running on Kubernetes!) and start collecting some data to build a pretty cool app.
Another example of how InfluxDB is a good monitoring tool for Kubernetes. This one is a specific implementation from the Influx folks, built with native Kubernetes operators and using Telegraf and the whole stack. We were pretty impressed with how complete and turnkey this solution looks, and eager to try it out. 🔑
If you already have a monitoring solution in place (which you hopefully do), how are you monitoring that system? Or, who watches the watchmen? Everything should be monitored, including your monitoring system. Adding Influx in here is a great idea and easy to integrate. 🕵🏼
This isn’t Influx related, but really wanted to make sure to share. The NSA has recently published a Kubernetes Hardening Guide. There’s nothing too surprising in here, but it’s a great resource to review your own clusters for best practices. And coming from the NSA, this is a sign of just how prevalent Kubernetes is becoming. 🛡
In case you missed it, the KubeCon NA schedule is live! This looks to be a great event with phenomenal content, whether you attend in person or virtually!