Shulou(Shulou.com)06/03 Report--

This article mainly explains "what are the health indicators of high-value Kubernetes". The content of the explanation is simple and clear, and it is easy to learn and understand. Please follow the editor's train of thought to study and learn "what are the health indicators of high-value Kubernetes".

Collapse cycle

A crash loop is the time that the Pod starts, crashes, and then continues to try to restart but cannot recover (it keeps crashing and restarting in a loop). When this happens, the application cannot run. This may be due to the crash of the application in Pod

It can also be caused by configuration errors in Pod or the deployment process, which makes debugging crash loops very tricky. When a crash cycle occurs, you need to know immediately what is happening and whether urgent measures need to be taken to maintain the application

The sequence is available.

CPU utilization

CPU utilization is only the number of CPU cycles used by the node. Monitoring is important for two reasons. First of all, you don't want to run out of processing resources for your application. If the application is limited by CPU, you need to increase the CPU allocation or add more sections to the cluster

Point. Second, you don't want CPU to be idle. If CPU usage remains low, it may be an overallocation of resources and a waste of expenditure.

Disk pressure

According to the threshold set in the Kubernetes configuration, disk pressure indicates that a node is using too much disk space or using disk space too fast. This is important for monitoring, because if the application legally requires more space, it may mean

More disk space needs to be added. Otherwise, it may mean that the application behaves abnormally. Either way, this situation needs attention.

Memory pressure

Memory stress is another resource condition that indicates that the node is out of memory. Similar to CPU resource configuration, you don't want to consume memory resources completely, but you don't want to over-allocate memory resources and waste costs. Then you need to pay attention to this situation, because it may mean that an application

There is a memory leak in the program.

PID pressure

PID stress is a rare situation in which the pod or container generates too many processes and prevents the node from getting the available process ID. Each node has a limited number of process ID, which is allocated among running processes; if the ID runs out

Other processes cannot be started Kubernetes allows you to set PID thresholds for Pod to limit its ability to perform runaway process generation, and PID stress conditions mean that one or more Pod has exhausted its assigned PID and needs to be checked.

The network is not available

All nodes need a network connection, and their status indicates whether there is a problem with the node's network connection. Either it is not set up correctly (due to routing exhaustion or misconfiguration), or there is a physical problem with the network connection to the hardware.

Job failed

The purpose of Job is to run pod for a limited amount of time and to disassemble it when the desired functionality is completed. If the Job fails to complete successfully due to a node crash or reboot or resource exhaustion, you need to know that the Job has failed. That's why we need to monitor job losses.

The reason for the defeat. They usually do not mean that the application is inaccessible, but if not resolved, they can cause problems.

Persistent volume failure

A persistent volume is a storage resource specified on the cluster and can be used as persistent storage for any Pod that requests it. During their lifecycle, they are bound to a container and then recycled when the container is no longer needed. If recycling fails for some reason, you need to

Be aware that there is a problem with persistent storage.

Delay time for suspending Pod

During the life cycle of pod, if it is waiting for scheduling on the node, its state is "pending". If the card is in a "suspended" state, it usually means that there are not enough resources to schedule and deploy Pod. Will need to update CPU and memory allocation, delete

Pod or add more nodes to the cluster.

Deployment failure

Deployment is used to manage stateless applications-Pod is interchangeable and does not need to be able to reach any particular single Pod, just a specific type of Pod. You need to pay close attention to the deployment to ensure that they are done correctly. The best way is to make sure to observe

The number of Deployment arrivals matches the number of Deployment required. If there is no match, one or more Deployment fails.

StatefulSets is not ready yet

StatefulSets is used to manage stateful applications, where Pod has a specific role and needs to reach other specific Pod. Instead of requiring only a specific type of Pod, as Deployment does. However, the monitoring is the same, and you need to ensure that what is observed

The number of StatefulSet matches the number of StatefulSet required. If there is a mismatch, one or more StatefulSet has failed.

DaemonSets is not ready.

DaemonSets is used to manage services or applications that need to run on all nodes in the cluster. If you have a log collection daemon or monitoring service to run on each node, you need to use DaemonSet. Monitoring is the same as Deployment: you need to make sure

The number of DaemonSet observed matches the number of DaemonSet required. If there is no match, one or more DaemonSet fails.

Thank you for your reading, the above is the content of "what are the health indicators of high-value Kubernetes". After the study of this article, I believe you have a deeper understanding of what the health indicators of high-value Kubernetes have, and the specific use needs to be verified in practice. Here is, the editor will push for you more related knowledge points of the article, welcome to follow!

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.