Exam 2022 Flashcards
(30 cards)
Prometheus is an open source monitoring tool. Prometheus server uses the following approach:
The correct answer is “pull”. Prometheus server uses a pull-based approach, where it periodically scrapes metrics from monitored targets, rather than using a push-based or event-driven approach.
An Amazon EC2 instance type defines:
The correct answer is “the hardware of the VM”. An Amazon EC2 instance type defines the hardware configuration of the virtual machine (VM), such as the number of virtual CPUs, amount of memory, and network performance, but does not directly define the amount of block storage in EBS or the software configuration of the VM.
Which of the following is not a supported CRI implementation of the lastest version of Kubernetes (>1.24)
- cri-o
-cri-dockerd
-dockershim
-containerd
dockershim
Docker shim was a component in Kubernetes for supporting Docker as the Container Runtime Interface (CRI). However, starting from Kubernetes version 1.20, Docker is no longer a supported runtime for Kubernetes. The Docker shim was deprecated in favor of other CRI implementations such as cri-o, cri-dockerd, and containerd.
Hence, dockershim is no longer a supported CRI implementation in the latest version of Kubernetes.
Which of the following is not a serverless compute platform (FaaS)?
- openFaaS
- Knative
- OpenWhisk
- AWS Lambda
- OpenStack
- OpenStack
A service registry is a database containing the network location of service instances. Which of the following is not a service registry tool?
- Netflix Eureka
- etcd
- istio
- istio
What is the deployment model of a private cloud?
- It is a part of the public cloud only accessible to your computer
- Resources in your own data center are offered via a network.
- You run a cloud infrastructure such as open stack on the resource of your own data center accessible to your own users.
- it is a private subnet in a data center of a cloud provider
- You run a cloud infrastructure such as open stack on the resource of your own data center accessible to your own users.
Which type of virtualization implementation is Docker based on?
- Function virtualization
- OS level virtualization
- Hosted virtualization
- Bare metal virtualization
- OS level virtualization
What is Blue/Green?
the green version is deployed alongside the blue version without any traffic being sent to the green version. After testing of the green version, the LB is updated to send traffic to the green version.
What is the main function of the jailer component in Firecracker?
- Limiting the available system calls for accessing the host kernel.
Which of the following K8s autoscalers is not valid?
- Cluster Autoscaler
- Vertical Pod Autoscaler
- Scheduled Pod Autoscaler
- Horizontal Pod Autoscaler
- Scheduled Pod Autoscaler
What is the purpose of command: docker rmi <image-name></image-name>
Delete an image from the local storage
Which service is used for:
- IP address and port only accessible inside the cluster:
- Allow to route traffic to systems outside the K8s cluster:
NodePort Service
LB service
ClusterIP Service
ExternalName Service
- IP address and port only accessible inside the cluster:
ClusterIP Service - Allow to route traffic to systems outside the K8s cluster:
ExternalName Service
- PID
- mount
- network
- cgroup
- user
What is Keystone in OpenStack?
Identity Management Service
(IMS)
What is Neutron in OpenStack?
- responsible for managing networks, ports, and attachments on infrastructure for virtual resources.
What does kubelet do?
Manages the lifecycle of every pod on the node
What is IoTCloud?
The sensors are connected via an IoT network to a gateway. Data are streamed to the cloud where an IoT platform stores the data and allows processing. Rule-based on AI-based analysis might lead to actuations which are sent to actuators via a gateway and the IoT network.
IoT Edge:
At the edge of the internet, edge servers are connected to the gateway and allow preprocessing of data from the sensors to reduce bandwidth and latency for actuations. The edge servers are mainly passively cooled with limited capabilities. IoT devices are connected through IoT Networks which can be short distance like bluetooth or wide area networks with low power characteristics like LoRa.
IoT Fog
Consists of a distributed continuum of scaled computing resources. Fog nodes are location aware and thus computation can be flexibly done at a suited location. Fog nodes can be managed as stand-alone resources or as a federated set of resources. The fog service models are the same as for the cloud: SaaS, PaaS, IaaS.
What is an example of IoT Fog?
Autonomous driving. Services might aggregate information from several cars over time and provide just-in-time low latency messages to the driver (like warnings). Fog is used for high computation so to predict how fast other cars will get to the free space or to integrate information from multiple edge systems.
Compare FaaS and IaaS on performance
This means that the performance of FaaS is optimized for short, stateless, and isolated function invocations, with automatic scaling and low latency.
IaaS can offer higher performance compared to FaaS for long-running, stateful, and resource-intensive workloads.
In summary, the performance of FaaS and IaaS depends on the use case and the workload being run.
Compare FaaS and IaaS on ease of application deployment
FaaS is designed to make it easy to deploy small, single-purpose functions that are triggered by events. This means that deploying an application on FaaS typically involves breaking down the application into individual functions, and deploying each function to the FaaS provider. This can simplify the deployment process and make it faster, as well as making it easier to scale and manage the application.
Deploying an application on IaaS involves setting up a virtual machine, configuring the operating system and application stack, and deploying the application.
Compare FaaS and IaaS on costs.
FaaS can reduce costs because you do not pay for idle VMs and do not run replicas for resilience.
This can be more cost-effective if you have long-running, stateful, and resource-intensive workloads, as you can reserve and pay for the resources you need in advance. However, if your resource usage is variable or unpredictable, you may end up paying for resources that are not being used, leading to higher costs compared to FaaS.
Vendor lock-in comparison of FaaS and IaaS
FaaS has more dependency to the service provider than IaaS. IaaS is less vendor dependent.
Additionally, proprietary APIs and services provided by the FaaS provider can further increase lock-in
