Our speaker is Alberto Castro. Alberto is in the process of transitioning from Software Development to DevOps, and as he likes to put it he will now get to “play with bigger blocks.” His interest range from software architecture, databases, cloud, serverless technologies, and now monitoring.
From the Prometheus web page: “Prometheus is an open-source systems monitoring and alerting toolkit originally built at SoundCloud.” It works natively with Docker and Kubernetes wich are instrumented to expose metrics that Prometheus can scrape. But Prometheus allows one to do a lot more. There are instrumentation libraries for Java, Pyton, and Go and with a system of exporters Prometheus can ingest metrics from databases, AWS services, linux and windows machines (bare metal or virtual), etc.
In this talk I will:
1. Explain the terminology.
2. Describe the use cases for Prometheus, what is it good for, and what is it not good for.
3. Describe in general terms the architecture of the system including the main monitoring tool, the alerting system and the storage database.
4. Introduce Grafana, which is the tool most used to display and graph Prometheus metrics.
5. Introduce PromQL the Prometheus query language.
5. Introduce the Node Exporter and BlackBox exporter.
In the Show and Tell section I will:
1. Install Prometheus and the 2 exporters described above in an EC2 instance and in another install Grafana.
2. Using Grafana create a few graphs to display the metrics.
These are a lot of topics so we will move fast, notes will be posted in Github after the talk.
Monthly Full Stack Meetup: