Wang Xiaowei , Tencent cloud container product manager , Keen to provide customers with efficient Kubernetes Usage mode , Reduce cost and increase efficiency for customers .
“ function “ The solution is whether the product has or does not have a capability , With “ function ” after , How to guide through good 、 Tips , And interaction to help users more “ Fluent ” The use of products is a problem worth thinking about .
for example ： LEGO has more than 3700 Two different modules , These modules can spell out an infinite number of models . However , The presentation of different models is completely different , It’s possible that just changing the position of a few small modules can bring unexpected results .
Efforts in recent years ,TKE A large number of new features have been proposed at the functional level ： for example ： adopt LB Direct connection Pod Lifting performance , Cluster upgrade in place , Virtual node , Dynamic scheduler , For more information Container service product trends .
Besides ,TKE Also pay attention to the overall user experience , I visited a number of companies for this purpose , Collected a lot of requirements , And according to the importance of users , Optimization and tool development , Users can take advantage of these optimization tips , Experience container services more smoothly , more “ Grounding gas ” Lower the threshold of product use , Enhance product experience .
colony / Node related
recommend ： Node search capability is enhanced
Node search support through Label/IP/ Whether to block / Node status And so on .
- Label Search format and native Kubernetes identical ：”k1=v1,k2=v2″
- IP Only precise query is supported in multi value search , single IP Search supports fuzzy queries
Usage mode ： Node management -> node -> Search for perhaps Node pool details page -> Drop down to the node list -> Search for
recommend ： Node initialization error display optimization
- 【CVM Shipment failed Detailed display of related errors 】 If because CVM Delivery failure leads to node purchase failure , Except for the newspaper LAUNCH_FAILED（ The delivery failed ）, It will also show the detailed reasons for the failure of delivery ： such as VPC subnet IP Insufficient / Insufficient disk resources / Lack of balance / Insufficient quota for placement groups wait .
- 【 Node initialization Detailed display of related errors 】 The node initialization failure steps and corresponding information will be displayed in detail in the node list ： such as Failed to hang up 、 Core component exception wait .
Usage mode ： Node management -> node -> Click View Status in the status bar to view the error
Node pool node operation standardization ： Support the blockade / deportation
Node operations in the node pool support blocking / deportation , Fully synchronized with the node list operation within the cluster , You can efficiently manage nodes in the node pool .
Usage mode ： Node pool details page -> Drop down to the node list -> Operate on a single node in the operation column perhaps Select the batch node and perform batch operation （ Click on the above 【 More operations 】 Buttons can be blocked in batches / Unblock node ）
Delete Protection Logic Optimization
Customer feedback ： If the delete by mistake option is turned on when creating a cluster , If creation fails after opening , Customers can’t delete .
Now you can go to 【 Cluster list page 】 Turn off delete Protection for the cluster .
Usage mode ： Cluster management -> Cluster operation list -> more -> Turn off cluster deletion protection
Public network IP Show optimization
Some customers reported that the public network IP The configuration entry of is too hidden , The default bandwidth value is too small , Unable to meet business needs .
We take a graphical form （ Drag bar ） Instead of the input box , Enhance the display effect .
Usage mode ： The new node / New node pool -> Choose a model -> Drag to the bottom -> Set the bandwidth of the public network
Node pool shows information structure / Advanced parameter display
- Add node pool core definition / Overview of functions , Let customers quickly understand the usage scenarios of node pool
- Structured display of node pool information （ Aggregate by module , It also shows more advanced parameters , Improved observability ）
- Node pool basic information ( Elastic scaling related configuration 、 The number of nodes )
- Node configuration details （ Node template related configuration 、 Node pool expansion node automatically inherited configuration ）
Nodes jump directly CVM Optimize
The old version of jump CVM The button display is not stable , And the entrance doesn’t stand out .
Usage mode ： Hover over the node ID It’s about ( If you click directly on the node ID, Will jump to Node Details page ) -> Click on “ Jump to CVM Example details page ” -> Will jump to CVM Example details page .
The container network is added when creating nodes IP check
Before you create a node , According to your container network configuration （ Network mode 、 Segments, etc ） To determine whether the container network can meet the requirements of creating these nodes . Considering some special scenarios, you may not need to verify the container network , We provide the parameters skip You can skip the check .（ For detailed description of parameters, please refer to API file ）
Usage mode ： The new node / Add existing nodes / New node pool -> More Settings -> Set up skip Parameters
When adding an existing node to a cluster, you can fill in the mount partition /LVM name
When adding an existing node, if you partition the disk /LVM, Please fill in the name of the partition you want to mount /LVM name , If the disk is not partitioned or LVM, No need to fill in the mount partition , Just fill in the mount point .
Usage mode ： Add existing nodes -> Select node -> Check data disk mount -> Fill in the mount area and mount point
Kubernetes Object related
recommend ： Batch import key value pairs ： Configuration Management & Environment variable optimization
Support importing batch key value pairs from the cutting board at one time . When the variable name is null , Paste one or more lines in the variable name key=value or key: value The key value pair of can realize fast batch input .
Usage mode 1： The workload -> New load -> Instance content -> environment variable
Usage mode 2： Configuration Management -> newly build ConfigMap/Secret -> Content
recommend ：CronJob Support to reserve quantity / Pause / Manual trigger / Show by Cronjob Generated Job list
- Keep success Job Count ： prevent Job Infinite generation , Impact on cluster performance
- Retention failed Job Count ： prevent Job Infinite generation , Impact on cluster performance
- Pause ： Will stop Cronjob Operation of , No longer will new Job. Support recovery
- Manual trigger ： No need to wait Cronjob Set up Crontab Time supports the use of Cronjob Inside Job The template generates a new Job
Usage mode ： The workload -> Cronjob -> newly build -> Keep success Job Count / Retention failed Job Count
Usage mode ： The workload -> Cronjob list -> more -> Manual trigger / Pause
Click on the image above CronJob The name of , You can see that the CronJob Generated, running Job list ：
HostAliases Used to Pod Of /etc/hosts Add an entry to the file , Can be in Pod Level override resolution of the host name . In order to increase Pod Of DNS To configure .
Usage mode ： The workload -> New load -> Advanced settings -> Host alias
lifecycle : The container’s life cycle event hook handler . The container can specify a handler for each event , When a container starts ,Kubernetes Will be sent immediately postStart event ; Before the container is terminated , Kubernetes Will send a preStop event .
Usage mode ： The workload -> New load -> Instance content -> Life cycle
Service Default “ Access is within the cluster only ”;Service Type explanation adds corresponding English nouns
- “ Access is within the cluster only ” The access mode is in line with the usage habit
- Comparing Chinese and English nouns to reduce the cost of understanding , accord with Kubernetes Original usage habit
Onekey “ Redeployment ” stay Deployment、StatefulSet and DaemonSet Upper support
and Deployment equally ,StatefulSet and DaemonSet Now it supports one click redeployment , Remove the trouble of reconfiguration after deletion .
Usage mode ： The workload -> Load list -> more -> Redeployment
Ingress Certificates support inheritance LB Certificate
Support direct use CLB Automatically create a certificate for Ingress Of HTTPS certificate , Save the tedious Secret Create a process .
Usage mode ： Services and routing -> Ingress -> newly build Ingress -> Monitor port check Https:443 -> New key -> Select the server certificate
Secret Support TLS type
except CLB certificate , You sometimes use your own certificate , Saving the certificate requires using TLS Type of Secret
Usage mode 1： Configuration Management -> Secret -> newly build Secret -> TLS certificate
Usage mode 2： Services and routing -> Ingress -> newly build Ingress -> Nginx Load Balancer -> Monitor port check Https:443 -> New key
subPathExpr Support configuration on the console
subPathExpr, Support will environment variable Mapping to Volume A subpath in
Usage mode ： New load -> Working with data volumes -> Add environment variables -> Use mount point -> choice subPathExpr
kube-system、kube-node-lease、kube-public The resource support in the console supports modification
In the past, in order to prevent users from touching , The console limits the modification of resources in these namespace , But in fact, some users themselves have the resource management rights of these namespace , So the new version removes these restrictions from the console , The authority logic is handed over to the background for judgment , If you don’t have permissions for the relevant namespace , Modifying resources will also report errors .
Usage mode ： New load -> Select a namespace -> Modify resource allocation
Usage mode ： Authorization management -> Choose any entry -> RBAC Policy generator -> Select a sub account -> Select a namespace / jurisdiction
Add multiple suspension tips
Add floating prompt to the field names of all resource list pages , Better for users with smaller screens
Add floating prompt to the log entry container name in the workload , Better for users with smaller screens
Operation and maintenance capability
recommend ： Log collection supports automatic extraction of regular expressions based on log samples
When you select single line full regular in log collection rules 、 When extracting patterns from multiline completely regular patterns , Regular expressions can be generated automatically according to the user’s own log samples , And automatically extract the corresponding key value pairs , There is no need for users to learn and configure regular expressions themselves
Usage mode ：1. Cluster operation and maintenance -> 2. Log rules -> 3. newly build -> 4. Enter the name of the log collection rule -> 5. Extraction method selection “ A single – Completely regular ” -> 6. Sample input log -> 7. Click regular expression to automatically generate -> 8. Move the cursor to select the string that needs to be regularized -> 9. Input key -> 10. confirm
recommend ： Log collection supports automatic verification of whether the log can be extracted normally according to the log sample
For user filled regular expressions 、 Separator, etc , Support verification according to user log sample , Automatically extract all value, To verify whether the log can be parsed normally , Avoid finding expression configuration errors when retrieving logs .
Usage mode ：1. Cluster operation and maintenance -> 2. Log rules -> 3. newly build -> 4. Enter the name of the log collection rule -> 5. Extraction method selection “ A single – Completely regular ” -> 6. Sample input log -> 7. Enter regular expressions manually -> 8. verification
Log collection supports multi line regular extraction mode
Support multiline regular parsing pattern （ Apply to java Programs and other multi line logs need to retrieve logs with key values ）
Be careful ： To use the multi row regular extraction mode, you need to contact the back end to upgrade the collection component version
Usage mode ： New log collection rule -> The second step is to select “ Multiple lines – Completely regular patterns ”
trusteeship Prometheus Support Singapore 、 Virginia and other places
Cloud native monitoring （ trusteeship Prometheus） It’s Singapore / Virginia and other places , Welcome to use .
Kubernetes It’s powerful , But the modules are complex , It costs a lot to start with .TKE Focus on the upgrade of function and experience , If you have any suggestions on function and experience , Welcome to comment on , We’re going to randomly pick three people to leave a message , Send a little cowboy around Tencent .
adopt LB Direct connection Pod Lifting performance : 【https://cloud.tencent.com/document/product/457/41897】
Cluster upgrade in place : 【https://cloud.tencent.com/document/product/457/32192】
Virtual node :【 https://cloud.tencent.com/document/product/457/53027】
Dynamic scheduler : 【https://cloud.tencent.com/document/product/457/50843】
Container service product trends : 【https://cloud.tencent.com/document/product/457/41088】
API file : 【https://cloud.tencent.com/document/product/457/36707】
Cloud native monitoring : 【https://cloud.tencent.com/document/product/457/49889】
Recommended in the past
- An article will take you to understand etcd 3.5 The core feature of
- On the cloud 618, Tencent cloud container helps 1.3 Fold up
- tencent TencentOS Ten years of cloud based iterative evolution
- kubernetes Standard Guide for cost reduction and efficiency enhancement | Resource utilization improvement tools
- stay TKE Use in Velero Migrating and replicating cluster resources
This article is from WeChat official account. –
Tencent cloud native （gh_faa30cb2ba5b）
, author ： Wang Xiaowei
The source and reprint of the original text are detailed in the text , If there is any infringement , Please contact the
Original publication time ：
Participation of this paper Tencent cloud media sharing plan , You are welcome to join us , share .