Kubernetes Engine Flashcards
Uptime Checks
Resource health is an integral part of cloud administration. Operations Suite can validate if a service is available by ensuring it can be accessed from any GCP data center globally. A best practice is to configure an uptime check, which is an alerting policy to ensure the system properly utilizes the default configuration. An uptime check can be created directly from the navigation pane in Operations Suite or using the Alerting Policy interface
How do engineers monitor Kubernetes?
Numerous monitoring tools can help an engineer ensure the system remains operational, including uptime checks and alerting policies. Uptime checks and alerting policies can be applied to a single instance within a cluster or to measure the performance of the entire cluster.
Alerting Policies
Alerting policies, similar to uptime checks, inform you if the performance of an instance within a cluster, an entire cluster, or a container breaches an expected performance threshold.
how do you monitor Node Pools
node pools. Given a cluster is made up of several nodes, it only makes sense to evaluate the performance of each one. Scroll to the bottom of the cluster details page and then select the link under the node pool you would like to evaluate. Upon clicking the link, you will find details about the configuration of each node, the metadata parameters, security configuration details, instance group details, and node-specific performance metrics
Objects
Objects are persistent entities within Google Kubernetes Engine. Kubernetes uses objects as a means to manage the operational state of a cluster. A Kubernetes object has three main purposes:
*It helps identify what containerized applications are running.
*It provides resource availability among those applications.
*It supports policy assignment around those applications.
Workloads
Workloads are the packaging of containerized, hardware-independent, isolated applications. Each container, whether it contains an application or batch job, is known as a workload.
Pods
Pods are single instances in a cluster where a process is actively operating. A Pod must contain at least one container.
Storage
Pod shared-storage volumes are available between one or more containers.
Services
Services in Kubernetes Engine enable many sets of Pod endpoints to become a single resource using configurable grouping options. The default design of a Service includes a stable cluster IP address that a client may use to contact other Pods using the same Service.
ReplicaSet
The purpose of the ReplicaSet is to ensure a stable set of Pods run perpetually. ReplicaSets are associated with fields, including a selector to set parameters such as the number of Pods that can be acquired, the number of replicas that can be maintained, and a template for establishing the creation of new replicas should new Pods be required.
StatefulSet
Pods that maintain unique, persistent identities as well as stable hostnames that Kubernetes Engine can support regardless of schedule are known as a StatefulSet. The use of StatefulSets is similar to Deployments; however, unique identifiers are available to each Po
DaemonSet
Making sure that all nodes run Pods is essential to container management. The use of a DaemonSet ensures that if Pods are added to a node, similar actions occur with a cluster.
Jobs
Jobs are another type of controller object in GKE. Unlike other controller objects, a Job will operate until completion instead of ceasing when it reaches a given state. Jobs are most often associated with the handling with finite tasks, although they are quite useful for complex operational activities such as computational and batch-oriented tasks.
Cluster
A cluster is the core object in the Google Kubernetes Engine that runs on top of the container application (see Figure 4-1). The cluster master is responsible for the Kubernetes control plane processes. A control plane process might include the API server, scheduler, and resource controllers. The cluster master life cycle occurs during the creation and deletion of a cluster.
Draw a zonal GKE cluster
Wat does each secton do.
Control plane
The control plane runs the control plane processes, including the Kubernetes API server, scheduler, and core resource controllers. The lifecycle of the control plane is managed by GKE when you create or delete a cluster.
Nodes
A cluster typically has one or more nodes, which are the worker machines that run your containerized applications and other workloads. The individual machines are Compute Engine VM instances that GKE creates on your behalf when you create a cluster.
What are regional clusters like?
In contrast to zonal clusters that have a single control plane in a single zone, regional clusters increase the availability of both a cluster’s control plane and its nodes by replicating them across multiple zones in a region.
If one zone in a region experiences an outage, the cluster’s control plane remains accessible as long as two replicas of the control plane remain available.
During cluster maintenance such as a cluster upgrade, only one replica of the control plane is unavailable at a time, and the cluster is still operational.
The control plane is replicated across three zones of a region. For node pools, you can manually specify the zone(s) in which the cluster’s node pools run or you can use the default configuration, which replicates each node pool across three zones of the control plane’s region. All zones must be within the same region as the cluster’s control plane.
Use regional clusters to run your production workloads, as they offer higher availability than zonal clusters.
