FAQ – Kubernetes


How to query logs from Loki using Grafana

As soon as you are logged into your Grafana, choose Explore from the menu on the left side. Then check “Loki” in the top left dropdown menu. This contains some of your log streams and examples for filtering etc. Generally speaking if you want to get a more in-depth look at how to write queries for Loki, feel free to read through the LogQL docs. Next write your query in the text field at the top. Alternatively you can use the label selector next to the text field. In order to execute the query, press the keys Shift+Enter or click on the “Run Query” button in the upper right corner.

What are node groups?

A node group describes a number of virtual machines that have various attributes as a group. Essentially, it determines which flavor – i.e. which VM model – will be used within this group. However, other attributes can also be selected. Each node group can be scaled vertically at any time independently of the others.

Why should I use node groups?

In short node groups are suitable for executing pods on specific nodes. For example, it is possible to define a group with the “Availability Zone” attribute. In addition to the already existing “default node group”, which distributes the nodes relatively at random across all availability zones, you can create additional node groups. These can be explicitly started in one availability zone. Within the Kubernetes cluster, you can even divide your pods into the corresponding availability zones or node groups.

What is Managed Kubernetes?

The easiest way to a functional Kubernetes cluster is certainly to use a managed Kubernetes offer. As a customer, you consume an optional, highly available Kubernetes API, which is ultimately used to operate the Kubernetes cluster. The provider then takes care of updates, availability and operation of the K8s cluster. Payment is subsequently made according to the cloud resources deployed and used. Hence the billing model only has marginal differences.

What is Kubernetes on-premises?

In total contrast, you can operate your Kubernetes yourself in the data center. In order to achieve a cloud-like functionality in your own data center, the managed Kubernetes solutions has to be reproduced. Nonetheless if you are lucky, you will already have some of the necessary components in operation. You might have some technical difficulties though, like automatic deployments, network and storage.

What is a pod?

A pod defines a group of containers (often just one) that share a common namespace on a host. The common namespaces (e.g. common file system or network) facilitate simple communication between the containers.You can always reach a pod through a unique IP in the cluster. Meanwhile, pods are only in use indirectly through a deployment.

How does Prometheus work?

In short the basic idea is that you start one or many Prometheus instances with the Prometheus Operator. Which in turn the ServiceMonitor dynamically configures. Usually this means that you can dock it to a normal Kubernetes service with a ServiceMonitor. In addition this also reads the endpoints and configures the corresponding Prometheus instance accordingly. If the service or the endpoints change however, for example in the number or the endpoints have new IPs, the ServiceMonitor detects this and reconfigures the Prometheus instance every time. Additionally, manual configurations are also carried out by using configmaps.