C_KYMD_01 SAP Certified Development Associate - Side-by-Side Extensibility based on SAP BTP, Kyma runtime Questions and Answers

• A DaemonSet ensures that all (or some) Nodes run a copy of a Pod. As nodes are added to the cluster, Pods are added to them. As nodes are removed from the cluster, those Pods are garbage collected. Deleting a DaemonSet will clean up the Pods it created2.
• Some typical uses of a DaemonSet are: running a cluster storage daemon on every node, running a logs collection daemon on every node, running a node monitoring daemon on every node2.
• DaemonSets are ideal in a variety of real-world use cases: Running Node monitoring agents, Running log collection agents, Running network plugins, Running service meshes3.

References:

• Kubernetes DaemonSet - What It is & How to Use It (Example)
• DaemonSet | Kubernetes
• Kubernetes Daemonset: A Practical Guide - Spot.io

kubectl is a command-line tool that allows you to run commands against Kubernetes clusters1. You can use kubectl in SAP BTP, Kyma runtime for the following purposes:

• To run interactive commands. kubectl supports various commands that let you interact with your cluster resources, such as creating, updating, deleting, describing, or executing them. For example, you can use kubectl create to create a resource from a file or kubectl exec to run a command in a container2.
• To view pod logs. kubectl allows you to view the logs of a pod or a container in a pod using the kubectl logs command. This can help you troubleshoot issues or monitor the pod activity3.
• To forward ports from remote machines to pods. kubectl enables you to forward one or more local ports to a pod using the kubectl port-forward command. This can allow you to access a pod from your local machine for debugging or testing purposes4.

References:

• 1: Kubernetes / Overview of kubectl
• 2: Kubernetes / kubectl Cheat Sheet
• 3: Kubernetes / Viewing logs
• 4: Kubernetes / Use Port Forwarding to Access Applications in a Cluster

The API Gateway is a component that enables exposing and securing services outside the Kyma cluster using the APIRule custom resource (CR)1. The API Gateway has the following characteristics:

• It is the central point of contact for all external traffic that enters the Kyma cluster. The API Gateway is a custom-configured Istio Ingress Gateway that is installed in the kyma-system namespace called kyma-gateway2. The kyma-gateway acts as a reverse proxy that routes the incoming requests to the appropriate services based on the host name and the path specified in the APIRule CR3.
• It uses Envoy Proxy to handle traffic and forward it to the correct service. The API Gateway is based on the Istio Gateway, which is a load balancer that operates at the edge of the mesh and enables ingress and egress traffic for the cluster. The Istio Gateway is implemented by a pod that runs an Envoy Proxy container, which is responsible for receiving and forwarding the traffic according to the Istio routing rules.

References:

• 1: Kyma / API Gateway v2
• 2: Using the API Gateway to Expose Services - SAP Learning
• 3: SAP BTP, Kyma Runtime API Gateway future architecture based on Istio | SAP Blogs
• [4]: Istio / Gateway
• [5]: Istio / Envoy

Stateless workloads are workloads that do not store any data or application state on the host that serves the requests. Instead, they rely on the information provided by the client or an external service, such as a database, to process each request independently. Stateless workloads do not keep any references to past transactions or interactions, and they can be handled by any available server. Stateless workloads are more scalable, fault-tolerant, and easier to manage than stateful workloads, which require persistent storage and session management. Some examples of stateless workloads are web servers, print servers, or microservices. References: Stateful vs stateless - Red Hat, Create stateless workloads | Google Distributed Cloud Hosted - Google Cloud, Stateless vs Stateful Kubernetes - WEKA

A zero downtime deployment is a technique that allows updating an application without interrupting its availability or functionality. It usually involves creating a separate environment for the new version and switching the traffic to it once it is ready. This way, the users do not experience any disruption or errors during the deployment process. A zero downtime deployment can be achieved by using different strategies, such as blue-green deployments, canary deployments, or rolling updates123. References: Zero Downtime Deployment Techniques: Rolling Update, Zero Downtime Deployment Techniques - Blue-Green Deployments, 5 Effective Deployment Strategies To Release Applications In 2021

SAP BTP, Kyma runtime provides a vendor-neutral integration point for traces, metrics, and logs, mainly using the OpenTelemetry Protocol (OTLP). For this option, the system you want to connect must support OTLP natively. The observability tools that support OTLP are Loki and Jaeger. Loki is a horizontally-scalable, highly-available, multi-tenant log aggregation system inspired by Prometheus. Jaeger is a distributed tracing system that monitors and troubleshoots transactions in complex distributed systems. Kiali and Grafana are not part of the observability architecture of SAP BTP, Kyma runtime, but they can be installed separately if needed. References: From Observability to Telemetry – a strategy shift in SAP BTP, Kyma Runtime, Accessing the built-in observability tools of Kyma runtime, SAP Help Portal - SAP Online Help

A pod in Kubernetes represents the smallest deployable units of computing that you can create and manage in Kubernetes. A pod is a group of one or more containers, with shared storage and network resources, and a specification for how to run the containers. A pod can run a single container or multiple containers that need to work together. Pods are the basic building blocks of Kubernetes applications, and they can be scaled, updated, and monitored by higher-level resources, such as deployments, services, or jobs. References: Pods | Kubernetes, What is a Kubernetes pod? - Red Hat, Kubernetes Pods - GeeksforGeeks

 A StatefulSet is a workload type that runs one or more pods that maintain a stable identity and persistent storage1. StatefulSets are suitable for applications that require one or more of the following2:

• Ordered, graceful deployment and scaling: StatefulSets ensure that pods are created and deleted in a sequential, ordered manner. Pods are assigned ordinal numbers that reflect their order in the set. Pods are scaled in reverse order of creation. StatefulSets also respect pod disruption budgets to limit the number of pods that are down simultaneously.
• Stable, persistent storage: StatefulSets use PersistentVolumeClaims to provide each pod with its own persistent storage. The PersistentVolumeClaim of a pod is not deleted when the pod is deleted, which preserves the state of the application across pod rescheduling.
• Stable, unique network identifiers: StatefulSets use a headless service to provide each pod with a unique network identity. Pods have a stable hostname based on their ordinal number and the name of the StatefulSet. Pods can use this hostname to communicate with each other and access their own or other pods’ data.

References: 1(https://kubernetes.io/docs/concepts/workloads/controllers/statefulset/),  2(https://kubernetes.io/docs/tutorials/stateful-application/basic-stateful-set/)

 The Kyma Eventing process uses the following technologies:

• HTTP-post requests: HTTP requests are delivered via POST-requests in terms of sending and receiving the events1.
• Jet Stream: Jet Stream provides the back end for the Kyma Eventing process1.
• NATS: NATS is an open source messaging system that is used for the Kyma Eventing process12. References: Discovering Kyma Eventing - SAP Learning, In-cluster eventing in SAP BTP, Kyma runtime | SAP Blogs

 SAP BTP, Kyma runtime uses Fluent Bit as the tool for log collection. Fluent Bit is an open-source log processor and forwarder that allows you to collect data and logs from different sources, enrich them with filters, and send them to multiple destinations. Fluent Bit is deployed as a DaemonSet on every node of the Kyma runtime cluster, and collects logs from the containers running on the node. Fluent Bit forwards the logs to Loki, a log aggregation system that is compatible with Prometheus12. References: Fluent Bit, Logging | SAP Help Portal.

To pull a container image from a private registry, you need to create a secret with the type “kubernetes.io/dockerconfigjson” and provide the credentials in the pod manifest via “imagePullSecrets” in spec.template.spec. This way, the kubelet can use the secret to authenticate with the private registry and pull the image. The secret type “Opaque” is not suitable for this purpose, as it is a generic type that can store any data. The pod manifest does not support providing credentials via “imagePullSecrets” in spec.template.metadata.annotations, as this is not a valid field. References: Side-by-Side Extensibility Based on SAP BTP, Kyma Runtime - Unit 3 - Lesson 2: Using Secrets, [Kubernetes Documentation - Pull an Image from a Private Registry]

A service manifest is a YAML file that defines a service, which is a Kubernetes resource that groups a set of pods and provides a way to access them within a cluster1. A service manifest must have the following mandatory metrics2:

• Kind: The kind of the resource, which must be Service for a service manifest.
• Type: The type of the service, which determines how the service is exposed. The possible values are ClusterIP, NodePort, LoadBalancer, or ExternalName3.
• Ports: The list of ports that the service exposes. Each port must have a name, a protocol, and a port number.

References:

• 1: Kubernetes / Services
• 2: Kubernetes / Service - v1 core
• 3: Kubernetes / Publishing Services (ServiceTypes)

SAP BTP, Kyma runtime uses the following open-source projects to expose and analyze in-cluster telemetry data:

• Loki: Loki is a log aggregation system that collects logs from various sources and stores them in a scalable and queryable way. Loki is integrated with Grafana to provide a unified interface for log analysis1.
• Prometheus: Prometheus is a time-series database that collects metrics from instrumented applications and stores them in a time-series database. Prometheus is a pull-based system, which means that applications need to expose an endpoint from which Prometheus can scrape the metrics2.
• Grafana: Grafana is a visualization and analysis tool to visualize and query time-series data. With Grafana, you can create dashboards that show the current state of your applications. Grafana can easily visualize data from Prometheus, Loki, and other sources2.

References:

• 1: Accessing the built-in observability tools of Kyma runtime
• 2: Describing Observability in Kyma

 A one off/non parallel job is a workload type that runs a single task to completion and then terminates. This is suitable for running a database migration on SAP BTP, Kyma runtime, as it does not require scaling, persistence, or parallelism. A parallel job, on the other hand, is a workload type that runs multiple tasks in parallel and then terminates. This is useful for batch processing or data analysis, but not for database migration. A StatefulSet is a workload type that runs one or more pods that maintain a stable identity and persistent storage. This is useful for stateful applications such as databases or message brokers, but not for database migration. A DaemonSet is a workload type that runs one pod on each node in the cluster. This is useful for providing node-level services such as logging or monitoring, but not for database migration. References: Workload Concepts, Kubernetes Workloads

 You can use the kubectl create -f command to create a Kubernetes object from a file. The file can be a YAML or JSON file that specifies the configuration of the object, such as its apiVersion, kind, metadata, and spec. The file can also be a URL that points to a remote file. The kubectl create -f command will send a POST request to the Kubernetes API server with the content of the file, and the API server will create the object according to the file. The other options are not valid kubectl commands. References: Imperative Management of Kubernetes Objects Using Configuration Files, Kubernetes Object Management