Discover the different options to scale your platform based on the traffic load you receive
When talking about Kubernetes, you’re always talking about the flexibility options that it provides. Usually, one of the topics that come into the discussion is the elasticity options that come with the platform — especially when working on a public cloud provider. But how can we really implement it?
Before we start to show how to scale our Kubernetes platform, we need to do a quick recap of the options that are available to us:
Cluster Autoscaler: When the load of the whole infrastructure reaches its peak, we can improve it by creating new worker nodes to host more service instances.
Horizontal Pod Autoscaling: When the load for a specific pod or set of pods reaches its peak, we deploy a new instance to ensure that we can have the global availability that we need.
Let’s see how we can implement these using one of the most popular Kubernetes-managed services, Amazon’s Elastic Kubernetes Services (EKS).
Setup
The first thing that we’re going to do is create a cluster with a single worker node to demonstrate the scalability behavior easily. And to do that, we’re going to use the command-line tool eksctl to manage an EKS cluster easily.
To be able to create the cluster, we’re going to do it with the following command:
After a few minutes, we will have our own Kubernetes cluster with a single node to deploy applications on top of it.
Now we’re going to create a sample application to generate load. We’re going to use TIBCO BusinessWorks Application Container Edition to generate a simple application. It will be a REST API that will execute a loop of 100,000 iterations acting as a counter and return a result.
BusinessWorks sample application to show the scalability options
And we will use the resources available in this GitHub repository:
GitHub – alexandrev/testeks
Contribute to alexandrev/testeks development by creating an account on GitHub.
We will build the container image and push it to a container registry. In my case, I am going to use my Amazon ECR instance to do so, and I will use the following commands:
With that, I can see and test the sample application using the browser, as shown below:
Swagger UI tester for the Kubernetes sample application
Horizontal pod autoscaling
Now, we need to start defining the autoscale rules, and we will start with the Horizontal Pod Autoscaler (HPA) rule. We will need to choose the resource that we would like to use to scale our pod. In this test, I will use the CPU utilization to do so, and I will use the following command:
That command will scale the replica set testeks from one (1) instance to five (5) instances when the CPU utilization percent is higher than 80%.
If now we check the status of the components, we will get something similar to the image below:
HPA rule definition for the application using CPU utilization as the key metric
If we check the TARGETS column, we will see this value: <unknown>/80%. That means that 80% is the target to trigger the new instances and the current usage is <unknown>.
We do not have anything deployed on the cluster to get the metrics for each of the pods. To solve that, we need to deploy the Metrics Server. To do so, we will follow the Amazon AWS documentation:
Installing the Kubernetes Metrics Server – Amazon EKS
The Kubernetes Metrics Server is an aggregator of resource usage data in your cluster, and it is not deployed by default in Amazon EKS clusters. For more information, see Kubernetes Metrics Server on GitHub. The Metrics Server is commonly used by other Kubernetes add ons, such as the
So, running the following command, we will have the Metrics Server installed.
And after doing that, if we check again, we can see that the current user has replaced the <unknown>:
Current resource utilization after installing the Metrics Server on the Kubernetes cluster
If that works, I am going to start sending requests using a Load Test inside the cluster. I will use the sample app defined below:
Auto Scaling Capacity with HPA – Ultimate Kubernetes Bootcamp
With Horizontal Pod Autoscaling, Kubernetes automatically scales the number of pods in a replication controller, deployment or replica set based on observed CPU utilization (or, with alpha support, on some other, application-provided metrics).
To deploy, we will use a YAML file with the following content:
And we will deploy it using the following command:
kubectl apply -f tester.yaml
After doing that, we will see that the current utilization is being increased. After a few seconds, it will start spinning new instances until it meets the maximum number of pods defined in the HPA rule.
Pods increasing when the load exceeds the target defined in previous steps.
Then, as soon as the load also decreases, the number of instances will be deleted.
Pods are deleted as soon as the load decreases.
Cluster autoscaling
Now, we need to see how we can implement the Cluster Autoscaler using EKS. We will use the information that Amazon provides:
Deployment edits that are needed to configure the Cluster Autoscaler
Now we need to run the following command:
kubectl -n kube-system set image deployment.apps/cluster-autoscaler cluster-autoscaler=eu.gcr.io/k8s-artifacts-prod/autoscaling/cluster-autoscaler:v1.17.4
The only thing that is left is to define the AutoScaling policy. To do that, we will use the AWS Services portal:
Enter into the EC service page on the region in which we have deployed the cluster.
Select the Auto Scaling Group options.
Select the Auto Scaling Group that has been created as part of the EKS cluster-creating process.
Go to the Automatic Scaling tab and click on the Add Policy button available.
Autoscaling policy option in the EC2 Service console
Then we should define the policy. We will use the Average CPU utilization as the metric and set the target value to 50%:
Autoscaling policy creation dialog
To validate the behavior, we will generate load using the tester as we did in the previous test and validate the node load using the following command:
kubectl top nodes
kubectl top nodes’ sample output
Now we deploy the tester again. As we already have it deployed in this cluster, we need to delete it first to deploy it again:
As soon as the load starts, new nodes are created, as shown in the image below:
kubectl top nodes showing how nodes have been scaled up
After the load finishes, we go back to the previous situation:
kubectl top nodes showing how nodes have been scaled down
Summary
In this article, we have shown how we can scale a Kubernetes cluster in a dynamic way both at the worker node level using the Cluster Autoscaler capability and at the pod level using the Horizontal Pod Autoscaler. That gives us all the options needed to create a truly elastic and flexible environment able to adapt to each moment’s needs with the most efficient approach.
Auto Scaling Capacity with HPA – Ultimate Kubernetes Bootcamp
With Horizontal Pod Autoscaling, Kubernetes automatically scales the number of pods in a replication controller, deployment or replica set based on observed CPU utilization (or, with alpha support, on some other, application-provided metrics).
One of the biggest announcements from the latest AWS re:Invent 2020 sessions was the release of EKS-D from Amazon. EKS-D is their open-source Kubernetes Distribution that’s now available for everyone to start using in their cloud provider or even on-premises.
It’s based on past findings and the entire process Amazon has undergone in managing their Kubernetes managed platform, Amazon EKS.
These announcements have many people asking themselves: “OK, I know Kubernetes, but what’s a Kubernetes distribution? And why should I care?”
So I’ll try to answer that with the knowledge I have, and I always try to use the same approach: a Kubernetes versus Linux model comparison.
Kubernetes is an open-source project, as you know, started by Google and is now being managed by the community and the Cloud Native Computing Foundation (CNCF), and you can find all the code available here:
GitHub – kubernetes/kubernetes: Production-Grade Container Scheduling and Management
Production-Grade Container Scheduling and Management – GitHub – kubernetes/kubernetes: Production-Grade Container Scheduling and Management
But let’s be honest: Not many of us are pulling that repo and trying to compile it to provide a cluster. That’s not how we usually work. If you follow the code path — downloading it, building it, and so on — this is usually named vanilla Kubernetes.
If we start with the Linux comparison, it’s the same situation as we have with the Linux kernel that most of the Linux distribution ships, but this is already compiled and available with a bunch of other tools all working together via the usual approach.
So that’s what a Kubernetes distribution is. They build Kubernetes. They provide other tools and components to enhance or provide more features and to focus on additional aspects like a security focus, a DevOps focus, or another focus. Another concept that usually is raised is the purity of distribution, and we try to talk about distribution that’s pure.
We call a distribution pure when it’s building Kubernetes, and that’s it. It leaves everything else to the developers or users to decide what they want to use on top of it.
What Are the Main Components Shipped in a Kubernetes Distribution?
The main components that can differ when we’re talking about a Kubernetes distribution are the following:
Container runtime and registries
We all know there’s more that one container runtime, and even if you weren’t aware of that, you’ve probably read all of the articles regarding the removal of Docker support in Kubernetes v1.20, as you can read in this awesome article from Edgar Rodriguez.
Kubernetes Just Deprecated Docker Support. What Now?
Will this kill Docker?
At this moment, it seems all runtimes should support the existing Container Runtime Interface, and runtimes like CRI-O, Containerd, or Kata seem to be the default options now.
Networking
Another topic that often differs when we’re talking about Kubernetes distributions is how they manage their network, and this is one of the most critical aspects of the whole platform.
As we have with the container runtime, a standard specification exists to cover that topic, and that’s the Container Network Interface (CNI). Several projects exist on this topic, like Flannel, Calico, Canal, and Wave. Also, some platforms provide their own component, like the Openshift SDN operator.
Storage
How to handle storage in Kubernetes is also very important, especially as we embrace this model in deployments that require stateful models. Different platforms can support different storage options, like file systems and so on.
Who Are the Top Players?
The first thing we need to be aware of is there are a huge number of Kubernetes distributions out there.
We’ll count the ones with a CNCF certification, and you can take a look at all of them here. At the moment of writing this article, we’re talking about 72 certified distributions.
These are the ones that I’d like to highlight today:
Red Hat OpenShift
The Red Hat OpenShift platform could be one of the most used platforms, especially in a private-cloud fashion. It could include most of the Red Hat services regarding storage, like GlusterFS and networking with OpenShift DNS. It has OKD as the open-source project that backs and contributes to the OpenShift platform. Check this article to see how to set up Openshift locally to test it
Mirantis
The former Docker enterprise that’s been acquired by Mirantis is another of the usual choices when we’re talking about supported platforms.
VMware Tanzu
VMware Tanzu, also coming from the acquisition of Pivotal from VMware, is a Kubernetes platform.
Canonical
Canonical (open source) is a platform from the company that develops and maintains Ubuntu. It’s another one of the important choices here and provides a variety of options, focusing not only on the common central mode but also on edge Kubernetes deployments with projects like MicroK8S and more options.
Rancher
Rancher (open source) is another one of the big players, focusing on following and extending the CNCF standards and also offering a big push for edge deployment with K3S. It also offers automated upgrades.
Summary
So, as you can see, the number of options available out there is huge. They all differ, so it’s important to take your time when you’re deciding your target platform based on your criteria for your project or your company.
And that’s without covering the managed platforms available out there that are becoming one of the more preferred options for companies so they can get all the flexibility from Kubernetes while not needing to handle the complexity of managing a Kubernetes platform themselves. But that’s a topic for another article — hopefully soon.
This article at least has provided you with more clarity about what a Kubernetes distribution is, the main differences among them, and a quick look at some of the key actors in this spectrum. Enjoy your day, and enjoy your life.
Find the greatest way to manage your Kubernetes development cluster
I need to start this article by admitting that I am an advocate of Graphical User Interfaces and everything that provides a way to speed up the way we do things and be more productive.
So when we talk about how to manage our Kubernetes cluster mainly for development purposes, you can imagine that I am one of those people who tries any available tool to make that journey easier. The ones who’ve started using Portainer to manage their local Docker engine or are a fan of the new dashboard in Docker for Windows/Mac. But that is far from reality.
In terms of Kubernetes management, I got used to typing all the commands to check the pods, the logs, the status of the cluster to do the port-forwards, etc. Any task I did was with a terminal, and I felt that it was the right thing to do. I did not even use a Kubernetes dashboard to have a web page for my Kubernetes environment. All of that changed last week when I met with a colleague who showed me what Lens could do.
Lens is a totally different story. I am not praising it because I am being paid to do so. This is an open source project that you can find on GitHub. But the way that it does the job is just awesome!
Image of Len showing the status of a Kubernetes cluster — Screenshot by the author.
The first thing I would like to mention regarding Lens is that it has multi-context support, so you can have all the different Kubernetes contexts available to switch following a Slack approach when we switch from different workspaces. It just reads your .kube/config file and makes all those contexts available to you to connect to the one you would like.
Kubernetes context selection in Lens
Once we have connected to one of these clusters, we have different options to see the status of it, but the first one is to check the Workloads using the Overview option:
Workloads Overview in Lens
Then, you can drill down to any pod or different object inside Kubernetes to check its status and at the same time do the main actions you usually do when you deal with a pod, such as check the logs, execute a terminal to one of the containers that belong to that pod, or even edit the YAML for that pod.
Pod options inside Lens
But Lens goes beyond the usual Kubernetes tasks because it also has a Helm integration, so you can check the releases that you have there, the version of the status, and so on:
Helm integration option in Lens
The experience of managing everything feels perfect. You are more productive as well. Even those who love the CLI and terminals need to admit that to do regular tasks, the Graphical approach and the mouse are faster than the keyboard — even for the defenders of the mechanical keyboard like myself.
So, I encourage you to download Lens and start using it right now. To do so, go to their main web page and download it:
Learn some tricks to analyze and optimize the usage that you are doing of the TSDB and save money on your cloud deployment.
In previous posts, we discussed how the storage layer worked for Prometheus and how effective it was. But in the current times, we are of cloud computing we know that each technical optimization is also a cost optimization as well and that is why we need to be very diligent about any option that we use regarding optimization.
We know that usually when we monitor using Prometheus we have so many exporters available at our disposal and also that each of them exposes a lot of very relevant metrics that we need to track everything we need to. But also, we should be aware that there are also metrics that we don’t need at this moment or we don’t plan to use it. So, if we are not planning to use, why do we want to waste disk space storing them?
So, let’s start taking a look at one of the exporters we have in our system. In my case, I would like to use a BusinessWorks Container Application that exposes metrics about its utilization. If you check their metrics endpoint you could see something like this:
# HELP jvm_info JVM version info # TYPE jvm_info gauge jvm_info{version="1.8.0_221-b27",vendor="Oracle Corporation",runtime="Java(TM) SE Runtime Environment",} 1.0 # HELP jvm_memory_bytes_used Used bytes of a given JVM memory area. # TYPE jvm_memory_bytes_used gauge jvm_memory_bytes_used{area="heap",} 1.0318492E8 jvm_memory_bytes_used{area="nonheap",} 1.52094712E8 # HELP jvm_memory_bytes_committed Committed (bytes) of a given JVM memory area. # TYPE jvm_memory_bytes_committed gauge jvm_memory_bytes_committed{area="heap",} 1.35266304E8 jvm_memory_bytes_committed{area="nonheap",} 1.71302912E8 # HELP jvm_memory_bytes_max Max (bytes) of a given JVM memory area. # TYPE jvm_memory_bytes_max gauge jvm_memory_bytes_max{area="heap",} 1.073741824E9 jvm_memory_bytes_max{area="nonheap",} -1.0 # HELP jvm_memory_bytes_init Initial bytes of a given JVM memory area. # TYPE jvm_memory_bytes_init gauge jvm_memory_bytes_init{area="heap",} 1.34217728E8 jvm_memory_bytes_init{area="nonheap",} 2555904.0 # HELP jvm_memory_pool_bytes_used Used bytes of a given JVM memory pool. # TYPE jvm_memory_pool_bytes_used gauge jvm_memory_pool_bytes_used{pool="Code Cache",} 3.3337536E7 jvm_memory_pool_bytes_used{pool="Metaspace",} 1.04914136E8 jvm_memory_pool_bytes_used{pool="Compressed Class Space",} 1.384304E7 jvm_memory_pool_bytes_used{pool="G1 Eden Space",} 3.3554432E7 jvm_memory_pool_bytes_used{pool="G1 Survivor Space",} 1048576.0 jvm_memory_pool_bytes_used{pool="G1 Old Gen",} 6.8581912E7 # HELP jvm_memory_pool_bytes_committed Committed bytes of a given JVM memory pool. # TYPE jvm_memory_pool_bytes_committed gauge jvm_memory_pool_bytes_committed{pool="Code Cache",} 3.3619968E7 jvm_memory_pool_bytes_committed{pool="Metaspace",} 1.19697408E8 jvm_memory_pool_bytes_committed{pool="Compressed Class Space",} 1.7985536E7 jvm_memory_pool_bytes_committed{pool="G1 Eden Space",} 4.6137344E7 jvm_memory_pool_bytes_committed{pool="G1 Survivor Space",} 1048576.0 jvm_memory_pool_bytes_committed{pool="G1 Old Gen",} 8.8080384E7 # HELP jvm_memory_pool_bytes_max Max bytes of a given JVM memory pool. # TYPE jvm_memory_pool_bytes_max gauge jvm_memory_pool_bytes_max{pool="Code Cache",} 2.5165824E8 jvm_memory_pool_bytes_max{pool="Metaspace",} -1.0 jvm_memory_pool_bytes_max{pool="Compressed Class Space",} 1.073741824E9 jvm_memory_pool_bytes_max{pool="G1 Eden Space",} -1.0 jvm_memory_pool_bytes_max{pool="G1 Survivor Space",} -1.0 jvm_memory_pool_bytes_max{pool="G1 Old Gen",} 1.073741824E9 # HELP jvm_memory_pool_bytes_init Initial bytes of a given JVM memory pool. # TYPE jvm_memory_pool_bytes_init gauge jvm_memory_pool_bytes_init{pool="Code Cache",} 2555904.0 jvm_memory_pool_bytes_init{pool="Metaspace",} 0.0 jvm_memory_pool_bytes_init{pool="Compressed Class Space",} 0.0 jvm_memory_pool_bytes_init{pool="G1 Eden Space",} 7340032.0 jvm_memory_pool_bytes_init{pool="G1 Survivor Space",} 0.0 jvm_memory_pool_bytes_init{pool="G1 Old Gen",} 1.26877696E8 # HELP jvm_buffer_pool_used_bytes Used bytes of a given JVM buffer pool. # TYPE jvm_buffer_pool_used_bytes gauge jvm_buffer_pool_used_bytes{pool="direct",} 148590.0 jvm_buffer_pool_used_bytes{pool="mapped",} 0.0 # HELP jvm_buffer_pool_capacity_bytes Bytes capacity of a given JVM buffer pool. # TYPE jvm_buffer_pool_capacity_bytes gauge jvm_buffer_pool_capacity_bytes{pool="direct",} 148590.0 jvm_buffer_pool_capacity_bytes{pool="mapped",} 0.0 # HELP jvm_buffer_pool_used_buffers Used buffers of a given JVM buffer pool. # TYPE jvm_buffer_pool_used_buffers gauge jvm_buffer_pool_used_buffers{pool="direct",} 19.0 jvm_buffer_pool_used_buffers{pool="mapped",} 0.0 # HELP jvm_classes_loaded The number of classes that are currently loaded in the JVM # TYPE jvm_classes_loaded gauge jvm_classes_loaded 16993.0 # HELP jvm_classes_loaded_total The total number of classes that have been loaded since the JVM has started execution # TYPE jvm_classes_loaded_total counter jvm_classes_loaded_total 17041.0 # HELP jvm_classes_unloaded_total The total number of classes that have been unloaded since the JVM has started execution # TYPE jvm_classes_unloaded_total counter jvm_classes_unloaded_total 48.0 # HELP bwce_activity_stats_list BWCE Activity Statictics list # TYPE bwce_activity_stats_list gauge # HELP bwce_activity_counter_list BWCE Activity related Counters list # TYPE bwce_activity_counter_list gauge # HELP all_activity_events_count BWCE All Activity Events count by State # TYPE all_activity_events_count counter all_activity_events_count{StateName="CANCELLED",} 0.0 all_activity_events_count{StateName="COMPLETED",} 0.0 all_activity_events_count{StateName="STARTED",} 0.0 all_activity_events_count{StateName="FAULTED",} 0.0 # HELP activity_events_count BWCE All Activity Events count by Process, Activity State # TYPE activity_events_count counter # HELP activity_total_evaltime_count BWCE Activity EvalTime by Process and Activity # TYPE activity_total_evaltime_count counter # HELP activity_total_duration_count BWCE Activity DurationTime by Process and Activity # TYPE activity_total_duration_count counter # HELP bwpartner_instance:total_request Total Request for the partner invocation which mapped from the activities # TYPE bwpartner_instance:total_request counter # HELP bwpartner_instance:total_duration_ms Total Duration for the partner invocation which mapped from the activities (execution or latency) # TYPE bwpartner_instance:total_duration_ms counter # HELP bwce_process_stats BWCE Process Statistics list # TYPE bwce_process_stats gauge # HELP bwce_process_counter_list BWCE Process related Counters list # TYPE bwce_process_counter_list gauge # HELP all_process_events_count BWCE All Process Events count by State # TYPE all_process_events_count counter all_process_events_count{StateName="CANCELLED",} 0.0 all_process_events_count{StateName="COMPLETED",} 0.0 all_process_events_count{StateName="STARTED",} 0.0 all_process_events_count{StateName="FAULTED",} 0.0 # HELP process_events_count BWCE Process Events count by Operation # TYPE process_events_count counter # HELP process_duration_seconds_total BWCE Process Events duration by Operation in seconds # TYPE process_duration_seconds_total counter # HELP process_duration_milliseconds_total BWCE Process Events duration by Operation in milliseconds # TYPE process_duration_milliseconds_total counter # HELP bwdefinitions:partner BWCE Process Events count by Operation # TYPE bwdefinitions:partner counter bwdefinitions:partner{ProcessName="t1.module.item.getTransactionData",ActivityName="FTLPublisher",ServiceName="GetCustomer360",OperationName="GetDataOperation",PartnerService="TransactionService",PartnerOperation="GetTransactionsOperation",Location="internal",PartnerMiddleware="MW",} 1.0 bwdefinitions:partner{ProcessName=" t1.module.item.auditProcess",ActivityName="KafkaSendMessage",ServiceName="GetCustomer360",OperationName="GetDataOperation",PartnerService="AuditService",PartnerOperation="AuditOperation",Location="internal",PartnerMiddleware="MW",} 1.0 bwdefinitions:partner{ProcessName="t1.module.item.getCustomerData",ActivityName="JMSRequestReply",ServiceName="GetCustomer360",OperationName="GetDataOperation",PartnerService="CustomerService",PartnerOperation="GetCustomerDetailsOperation",Location="internal",PartnerMiddleware="MW",} 1.0 # HELP bwdefinitions:binding BW Design Time Repository - binding/transport definition # TYPE bwdefinitions:binding counter bwdefinitions:binding{ServiceName="GetCustomer360",OperationName="GetDataOperation",ServiceInterface="GetCustomer360:GetDataOperation",Binding="/customer",Transport="HTTP",} 1.0 # HELP bwdefinitions:service BW Design Time Repository - Service definition # TYPE bwdefinitions:service counter bwdefinitions:service{ProcessName="t1.module.sub.item.getCustomerData",ServiceName="GetCustomer360",OperationName="GetDataOperation",ServiceInstance="GetCustomer360:GetDataOperation",} 1.0 bwdefinitions:service{ProcessName="t1.module.sub.item.auditProcess",ServiceName="GetCustomer360",OperationName="GetDataOperation",ServiceInstance="GetCustomer360:GetDataOperation",} 1.0 bwdefinitions:service{ProcessName="t1.module.sub.orchestratorSubFlow",ServiceName="GetCustomer360",OperationName="GetDataOperation",ServiceInstance="GetCustomer360:GetDataOperation",} 1.0 bwdefinitions:service{ProcessName="t1.module.Process",ServiceName="GetCustomer360",OperationName="GetDataOperation",ServiceInstance="GetCustomer360:GetDataOperation",} 1.0 # HELP bwdefinitions:gateway BW Design Time Repository - Gateway definition # TYPE bwdefinitions:gateway counter bwdefinitions:gateway{ServiceName="GetCustomer360",OperationName="GetDataOperation",ServiceInstance="GetCustomer360:GetDataOperation",Endpoint="bwce-demo-mon-orchestrator-bwce",InteractionType="ISTIO",} 1.0 # HELP process_cpu_seconds_total Total user and system CPU time spent in seconds. # TYPE process_cpu_seconds_total counter process_cpu_seconds_total 1956.86 # HELP process_start_time_seconds Start time of the process since unix epoch in seconds. # TYPE process_start_time_seconds gauge process_start_time_seconds 1.604712447107E9 # HELP process_open_fds Number of open file descriptors. # TYPE process_open_fds gauge process_open_fds 763.0 # HELP process_max_fds Maximum number of open file descriptors. # TYPE process_max_fds gauge process_max_fds 1048576.0 # HELP process_virtual_memory_bytes Virtual memory size in bytes. # TYPE process_virtual_memory_bytes gauge process_virtual_memory_bytes 3.046207488E9 # HELP process_resident_memory_bytes Resident memory size in bytes. # TYPE process_resident_memory_bytes gauge process_resident_memory_bytes 4.2151936E8 # HELP jvm_gc_collection_seconds Time spent in a given JVM garbage collector in seconds. # TYPE jvm_gc_collection_seconds summary jvm_gc_collection_seconds_count{gc="G1 Young Generation",} 540.0 jvm_gc_collection_seconds_sum{gc="G1 Young Generation",} 4.754 jvm_gc_collection_seconds_count{gc="G1 Old Generation",} 2.0 jvm_gc_collection_seconds_sum{gc="G1 Old Generation",} 0.563 # HELP jvm_threads_current Current thread count of a JVM # TYPE jvm_threads_current gauge jvm_threads_current 98.0 # HELP jvm_threads_daemon Daemon thread count of a JVM # TYPE jvm_threads_daemon gauge jvm_threads_daemon 43.0 # HELP jvm_threads_peak Peak thread count of a JVM # TYPE jvm_threads_peak gauge jvm_threads_peak 98.0 # HELP jvm_threads_started_total Started thread count of a JVM # TYPE jvm_threads_started_total counter jvm_threads_started_total 109.0 # HELP jvm_threads_deadlocked Cycles of JVM-threads that are in deadlock waiting to acquire object monitors or ownable synchronizers # TYPE jvm_threads_deadlocked gauge jvm_threads_deadlocked 0.0 # HELP jvm_threads_deadlocked_monitor Cycles of JVM-threads that are in deadlock waiting to acquire object monitors # TYPE jvm_threads_deadlocked_monitor gauge jvm_threads_deadlocked_monitor 0.0
As you can see a lot of metrics but I have to be honest I am not using most of them in my dashboards and to generate my alerts. I can use the metrics regarding the application performance for each of the BusinessWorks process and its activities, also the JVM memory performance and number of threads but things like how the JVM GC is working for each of the layers of the JVM (G1 Young Generation, G1 Old Generation) I’m not using them at all.
So, If I show the same metric endpoint highlighting the things that I am not using it would be something like this:
# HELP jvm_info JVM version info # TYPE jvm_info gauge jvm_info{version="1.8.0_221-b27",vendor="Oracle Corporation",runtime="Java(TM) SE Runtime Environment",} 1.0 # HELP jvm_memory_bytes_used Used bytes of a given JVM memory area. # TYPE jvm_memory_bytes_used gauge jvm_memory_bytes_used{area="heap",} 1.0318492E8 jvm_memory_bytes_used{area="nonheap",} 1.52094712E8 # HELP jvm_memory_bytes_committed Committed (bytes) of a given JVM memory area. # TYPE jvm_memory_bytes_committed gauge jvm_memory_bytes_committed{area="heap",} 1.35266304E8 jvm_memory_bytes_committed{area="nonheap",} 1.71302912E8 # HELP jvm_memory_bytes_max Max (bytes) of a given JVM memory area. # TYPE jvm_memory_bytes_max gauge jvm_memory_bytes_max{area="heap",} 1.073741824E9 jvm_memory_bytes_max{area="nonheap",} -1.0 # HELP jvm_memory_bytes_init Initial bytes of a given JVM memory area. # TYPE jvm_memory_bytes_init gauge jvm_memory_bytes_init{area="heap",} 1.34217728E8 jvm_memory_bytes_init{area="nonheap",} 2555904.0 # HELP jvm_memory_pool_bytes_used Used bytes of a given JVM memory pool. # TYPE jvm_memory_pool_bytes_used gauge jvm_memory_pool_bytes_used{pool="Code Cache",} 3.3337536E7 jvm_memory_pool_bytes_used{pool="Metaspace",} 1.04914136E8 jvm_memory_pool_bytes_used{pool="Compressed Class Space",} 1.384304E7 jvm_memory_pool_bytes_used{pool="G1 Eden Space",} 3.3554432E7 jvm_memory_pool_bytes_used{pool="G1 Survivor Space",} 1048576.0 jvm_memory_pool_bytes_used{pool="G1 Old Gen",} 6.8581912E7 # HELP jvm_memory_pool_bytes_committed Committed bytes of a given JVM memory pool. # TYPE jvm_memory_pool_bytes_committed gauge jvm_memory_pool_bytes_committed{pool="Code Cache",} 3.3619968E7 jvm_memory_pool_bytes_committed{pool="Metaspace",} 1.19697408E8 jvm_memory_pool_bytes_committed{pool="Compressed Class Space",} 1.7985536E7 jvm_memory_pool_bytes_committed{pool="G1 Eden Space",} 4.6137344E7 jvm_memory_pool_bytes_committed{pool="G1 Survivor Space",} 1048576.0 jvm_memory_pool_bytes_committed{pool="G1 Old Gen",} 8.8080384E7 # HELP jvm_memory_pool_bytes_max Max bytes of a given JVM memory pool. # TYPE jvm_memory_pool_bytes_max gauge jvm_memory_pool_bytes_max{pool="Code Cache",} 2.5165824E8 jvm_memory_pool_bytes_max{pool="Metaspace",} -1.0 jvm_memory_pool_bytes_max{pool="Compressed Class Space",} 1.073741824E9 jvm_memory_pool_bytes_max{pool="G1 Eden Space",} -1.0 jvm_memory_pool_bytes_max{pool="G1 Survivor Space",} -1.0 jvm_memory_pool_bytes_max{pool="G1 Old Gen",} 1.073741824E9 # HELP jvm_memory_pool_bytes_init Initial bytes of a given JVM memory pool. # TYPE jvm_memory_pool_bytes_init gauge jvm_memory_pool_bytes_init{pool="Code Cache",} 2555904.0 jvm_memory_pool_bytes_init{pool="Metaspace",} 0.0 jvm_memory_pool_bytes_init{pool="Compressed Class Space",} 0.0 jvm_memory_pool_bytes_init{pool="G1 Eden Space",} 7340032.0 jvm_memory_pool_bytes_init{pool="G1 Survivor Space",} 0.0 jvm_memory_pool_bytes_init{pool="G1 Old Gen",} 1.26877696E8 # HELP jvm_buffer_pool_used_bytes Used bytes of a given JVM buffer pool. # TYPE jvm_buffer_pool_used_bytes gauge jvm_buffer_pool_used_bytes{pool="direct",} 148590.0 jvm_buffer_pool_used_bytes{pool="mapped",} 0.0 # HELP jvm_buffer_pool_capacity_bytes Bytes capacity of a given JVM buffer pool. # TYPE jvm_buffer_pool_capacity_bytes gauge jvm_buffer_pool_capacity_bytes{pool="direct",} 148590.0 jvm_buffer_pool_capacity_bytes{pool="mapped",} 0.0 # HELP jvm_buffer_pool_used_buffers Used buffers of a given JVM buffer pool. # TYPE jvm_buffer_pool_used_buffers gauge jvm_buffer_pool_used_buffers{pool="direct",} 19.0 jvm_buffer_pool_used_buffers{pool="mapped",} 0.0 # HELP jvm_classes_loaded The number of classes that are currently loaded in the JVM # TYPE jvm_classes_loaded gauge jvm_classes_loaded 16993.0 # HELP jvm_classes_loaded_total The total number of classes that have been loaded since the JVM has started execution # TYPE jvm_classes_loaded_total counter jvm_classes_loaded_total 17041.0 # HELP jvm_classes_unloaded_total The total number of classes that have been unloaded since the JVM has started execution # TYPE jvm_classes_unloaded_total counter jvm_classes_unloaded_total 48.0 # HELP bwce_activity_stats_list BWCE Activity Statictics list # TYPE bwce_activity_stats_list gauge # HELP bwce_activity_counter_list BWCE Activity related Counters list # TYPE bwce_activity_counter_list gauge # HELP all_activity_events_count BWCE All Activity Events count by State # TYPE all_activity_events_count counter all_activity_events_count{StateName="CANCELLED",} 0.0 all_activity_events_count{StateName="COMPLETED",} 0.0 all_activity_events_count{StateName="STARTED",} 0.0 all_activity_events_count{StateName="FAULTED",} 0.0 # HELP activity_events_count BWCE All Activity Events count by Process, Activity State # TYPE activity_events_count counter # HELP activity_total_evaltime_count BWCE Activity EvalTime by Process and Activity # TYPE activity_total_evaltime_count counter # HELP activity_total_duration_count BWCE Activity DurationTime by Process and Activity # TYPE activity_total_duration_count counter # HELP bwpartner_instance:total_request Total Request for the partner invocation which mapped from the activities # TYPE bwpartner_instance:total_request counter # HELP bwpartner_instance:total_duration_ms Total Duration for the partner invocation which mapped from the activities (execution or latency) # TYPE bwpartner_instance:total_duration_ms counter # HELP bwce_process_stats BWCE Process Statistics list # TYPE bwce_process_stats gauge # HELP bwce_process_counter_list BWCE Process related Counters list # TYPE bwce_process_counter_list gauge # HELP all_process_events_count BWCE All Process Events count by State # TYPE all_process_events_count counter all_process_events_count{StateName="CANCELLED",} 0.0 all_process_events_count{StateName="COMPLETED",} 0.0 all_process_events_count{StateName="STARTED",} 0.0 all_process_events_count{StateName="FAULTED",} 0.0 # HELP process_events_count BWCE Process Events count by Operation # TYPE process_events_count counter # HELP process_duration_seconds_total BWCE Process Events duration by Operation in seconds # TYPE process_duration_seconds_total counter # HELP process_duration_milliseconds_total BWCE Process Events duration by Operation in milliseconds # TYPE process_duration_milliseconds_total counter # HELP bwdefinitions:partner BWCE Process Events count by Operation # TYPE bwdefinitions:partner counter bwdefinitions:partner{ProcessName="t1.module.item.getTransactionData",ActivityName="FTLPublisher",ServiceName="GetCustomer360",OperationName="GetDataOperation",PartnerService="TransactionService",PartnerOperation="GetTransactionsOperation",Location="internal",PartnerMiddleware="MW",} 1.0 bwdefinitions:partner{ProcessName=" t1.module.item.auditProcess",ActivityName="KafkaSendMessage",ServiceName="GetCustomer360",OperationName="GetDataOperation",PartnerService="AuditService",PartnerOperation="AuditOperation",Location="internal",PartnerMiddleware="MW",} 1.0 bwdefinitions:partner{ProcessName="t1.module.item.getCustomerData",ActivityName="JMSRequestReply",ServiceName="GetCustomer360",OperationName="GetDataOperation",PartnerService="CustomerService",PartnerOperation="GetCustomerDetailsOperation",Location="internal",PartnerMiddleware="MW",} 1.0 # HELP bwdefinitions:binding BW Design Time Repository - binding/transport definition # TYPE bwdefinitions:binding counter bwdefinitions:binding{ServiceName="GetCustomer360",OperationName="GetDataOperation",ServiceInterface="GetCustomer360:GetDataOperation",Binding="/customer",Transport="HTTP",} 1.0 # HELP bwdefinitions:service BW Design Time Repository - Service definition # TYPE bwdefinitions:service counter bwdefinitions:service{ProcessName="t1.module.sub.item.getCustomerData",ServiceName="GetCustomer360",OperationName="GetDataOperation",ServiceInstance="GetCustomer360:GetDataOperation",} 1.0 bwdefinitions:service{ProcessName="t1.module.sub.item.auditProcess",ServiceName="GetCustomer360",OperationName="GetDataOperation",ServiceInstance="GetCustomer360:GetDataOperation",} 1.0 bwdefinitions:service{ProcessName="t1.module.sub.orchestratorSubFlow",ServiceName="GetCustomer360",OperationName="GetDataOperation",ServiceInstance="GetCustomer360:GetDataOperation",} 1.0 bwdefinitions:service{ProcessName="t1.module.Process",ServiceName="GetCustomer360",OperationName="GetDataOperation",ServiceInstance="GetCustomer360:GetDataOperation",} 1.0 # HELP bwdefinitions:gateway BW Design Time Repository - Gateway definition # TYPE bwdefinitions:gateway counter bwdefinitions:gateway{ServiceName="GetCustomer360",OperationName="GetDataOperation",ServiceInstance="GetCustomer360:GetDataOperation",Endpoint="bwce-demo-mon-orchestrator-bwce",InteractionType="ISTIO",} 1.0 # HELP process_cpu_seconds_total Total user and system CPU time spent in seconds. # TYPE process_cpu_seconds_total counter process_cpu_seconds_total 1956.86 # HELP process_start_time_seconds Start time of the process since unix epoch in seconds. # TYPE process_start_time_seconds gauge process_start_time_seconds 1.604712447107E9 # HELP process_open_fds Number of open file descriptors. # TYPE process_open_fds gauge process_open_fds 763.0 # HELP process_max_fds Maximum number of open file descriptors. # TYPE process_max_fds gauge process_max_fds 1048576.0 # HELP process_virtual_memory_bytes Virtual memory size in bytes. # TYPE process_virtual_memory_bytes gauge process_virtual_memory_bytes 3.046207488E9 # HELP process_resident_memory_bytes Resident memory size in bytes. # TYPE process_resident_memory_bytes gauge process_resident_memory_bytes 4.2151936E8 # HELP jvm_gc_collection_seconds Time spent in a given JVM garbage collector in seconds. # TYPE jvm_gc_collection_seconds summary jvm_gc_collection_seconds_count{gc="G1 Young Generation",} 540.0 jvm_gc_collection_seconds_sum{gc="G1 Young Generation",} 4.754 jvm_gc_collection_seconds_count{gc="G1 Old Generation",} 2.0 jvm_gc_collection_seconds_sum{gc="G1 Old Generation",} 0.563 # HELP jvm_threads_current Current thread count of a JVM # TYPE jvm_threads_current gauge jvm_threads_current 98.0 # HELP jvm_threads_daemon Daemon thread count of a JVM # TYPE jvm_threads_daemon gauge jvm_threads_daemon 43.0 # HELP jvm_threads_peak Peak thread count of a JVM # TYPE jvm_threads_peak gauge jvm_threads_peak 98.0 # HELP jvm_threads_started_total Started thread count of a JVM # TYPE jvm_threads_started_total counter jvm_threads_started_total 109.0 # HELP jvm_threads_deadlocked Cycles of JVM-threads that are in deadlock waiting to acquire object monitors or ownable synchronizers # TYPE jvm_threads_deadlocked gauge jvm_threads_deadlocked 0.0 # HELP jvm_threads_deadlocked_monitor Cycles of JVM-threads that are in deadlock waiting to acquire object monitors # TYPE jvm_threads_deadlocked_monitor gauge jvm_threads_deadlocked_monitor 0.0
So, it can be a 50% of the metric endpoint response the part that I’m not using, so, why I am using disk space that I am paying for to storing it? And this is just for a “critical exporter”, one that I try to use as much information as possible, but think about how many exporters do you have and how much information you use for each of them.
Ok, so now the purpose and the motivation of this post are clear, but what we can do about it?
Discovering the REST API
Prometheus has an awesome REST API to expose all the information that you can wish about. If you have ever use the Graphical Interface for Prometheus (shown below) you are using the REST API because this is why is behind it.
Target view of the Prometheus Graphical Interface
We have all the documentation regarding the REST API in the Prometheus official documentation:
But what is this API providing us in terms of the time-series database TSDB that Prometheus is using?
TSDB Admin APIs
We have a specific API to manage the performance of the TSDB database but in order to be able to use it, we need to enable the Admin API. And that is done by providing the following flag where we are launching the Prometheus server --web.enable-admin-api.
If we are using the Prometheus Operator Helm Chart to deploy this we need to use the following item in our values.yaml
## EnableAdminAPI enables Prometheus the administrative HTTP API which includes functionality such as deleting time series. ## This is disabled by default. ## ref: https://prometheus.io/docs/prometheus/latest/querying/api/#tsdb-admin-apis ## enableAdminAPI: true
We have a lot of options enable when we enable this administrative API but today we are going to focus on a single REST operation that is the “stats”. This is the only method related to TSDB that it doesn’t require to enable the Admin API. This operation, as we can read in the Prometheus documentation, returns the following items:
headStats: This provides the following data about the head block of the TSDB:
numSeries: The number of series.
chunkCount: The number of chunks.
minTime: The current minimum timestamp in milliseconds.
maxTime: The current maximum timestamp in milliseconds.
seriesCountByMetricName: This will provide a list of metrics names and their series count.
labelValueCountByLabelName: This will provide a list of the label names and their value count.
memoryInBytesByLabelName This will provide a list of the label names and memory used in bytes. Memory usage is calculated by adding the length of all values for a given label name.
seriesCountByLabelPair This will provide a list of label value pairs and their series count.
To access to that API we need to hit the following endpoint:
GET /api/v1/status/tsdb
So, when I am doing that in my Prometheus deployment I get something similar to this:
We can also check the same information if we use the new and experimental React User Interface on the following endpoint:
/new/tsdb-status
Graphical Visualization of top 10 series count by metric name in the new Prometheus UI
So, with that, you will get the Top 10 series and labels that are inside your time-series database, so in case, some of them are not useful you can just get rid of them using the normal approaches to drop a series or a label. This is great, but what if all the ones shown here are relevant, what can we do about it?
Mmmm, maybe we can use PromQL to monitor this (dogfodding approach). So if we would like to extract the same information but using PromQL we can do it with the following query:
topk(10, count by (__name__)({__name__=~".+"}))
Top 10 of metric series generated and stored in the time series database
And now we have all the power at my hands. For example, let’s take a look not at the 10 more relevant but the 100 more relevants or any other filter that we need to apply. For example, let’s see the metrics regarding with the JVM that we discussed at the beginning. And we will do that with the following PromQL query:
topk(100, count by (__name__)({__name__=~"jvm.+"}))
Top 100 of metric series regarding to JVM metrics
So we can see that we have at least 150 series regarding to metrics that I am not using at all. But let’s do it even better, let’s take a look at the same but group by job names:
topk(10, count by (job,__name__)({__name__=~".+"}))
Result of checking the top 10 metric series count with the job that is generating them
Prometheus is one of the key systems in nowadays cloud architectures. The second graduate project from the Cloud Native Computing Foundation (CNCF) after Kubernetes itself, and is the monitoring solution for excellence in most of the workloads running on Kubernetes.
If you already have used Prometheus for some time, you know that it relies on a Time series database so Prometheus storage is one of the key elements. Based on their own words from the Prometheus official page:
Storage | Prometheus
An open-source monitoring system with a dimensional data model, flexible query language, efficient time series database and modern alerting approach.
Every time series is uniquely identified by its metric name and optional key-value pairs called labels, and that series is similar to the tables in a relational model. And inside each of those series, we have samples that are similar to the tuples. And each of the samples contains a float value and a milliseconds-precision timestamp.
Default on-disk approach
By default, Prometheus uses a local-storage approach storing all those samples on disk. This data is distributed in different files and folders to group different chunks of data.
So, we have folders to create those groups, and by default, they are a two-hour block and can contain one or more files depending on the amount of data ingested in that period of time as each folder contains all the samples for that specific timeline.
Additionally, each folder also has some kind of metadata files that help locate each of the data files’ metrics.
A file is persistent in a complete manner when the block is over, and before that, it keeps in memory and uses a write-ahead log technical to recover the data in case of a crash of the Prometheus server.
So, at a high-level view, the directory structure of a Prometheus server’s data directory will look something like this:
Remote Storage Integration
Default on-disk storage is good and has some limitations in terms of scalability and durability, even considering the performance improvement of the latest version of the TSDB. So, if we’d like to explore other options to store this data, Prometheus provides a way to integrate with remote storage locations.
It provides an API that allows writing samples that are being ingested into a remote URL and, at the same time, be able to read back sample data for that remote URL as shown in the picture below:
As always in anything related to Prometheus, the number of adapters created using this pattern is huge, and it can be seen in the following link in detail:
Integrations | Prometheus
An open-source monitoring system with a dimensional data model, flexible query language, efficient time series database and modern alerting approach.
Summary
Knowing how prometheus storage works is critical to understand how we can optimize their usage to improve the performance of our monitoring solution and provide a cost-efficient deployment.
In the following posts, we’re going to cover how we can optimize the usage of this storage layer, making sure that only the metrics and samples that are important to use are being stored, and also how to analyze which metrics are the ones used most of the time-series database to be able to take good decision about which metrics should be dropped and which ones should be kept.
In the previous post of these series regarding how to set up a Hybrid EKS cluster making use of both traditional EC2 machines but also serverless options using Fargate, we were able to create the EKS cluster with both deployment fashion available. If you didn’t take a look at it yet, do it now!
At that point, we have an empty cluster with everything ready to deploy new workloads, but we still need to configure a few things before doing the deployment. First thing is to decide which workloads are going to be deployed using the serverless option and which ones will use the traditional EC2 option.
By default, all the workloads deployed on the namespaces default and kube-system as you can see in the picture below form the AWS Console:
So that means that all workloads from the default namespace and the kube-system namespace will be deployed in a serverless fashion. If that’s what you’d like perfect. But sometimes you’d like to start with a delimited set of namespaces where you’d like to use the serverless option and rely on the traditional deployment.
We can check that same information using eksctl and to do that we need to type the following command:
eksctl get fargateprofile --cluster cluster-test-hybrid -o yaml
The output of that command should be something similar of the information that we can see in the AWS Console:
NOTE: If you don’t remember the name of your cluster you just need to type the command eksctl get clusters
So, this is what we’re going to do and to do that the first thing we need to do is to create a new namespace named “serverless” that is going to hold our serverless deployment and to do that we use a kubectl command as follows:
kubectl create namespace serverless
And now, we just need to create a new fargate profile that is going to replace the one that we have at the moment and to do that we need to make use again of eksctl to handle that job:
NOTE: We also can use not only namespace to limit the scope of our serverless deployment but also tags, so we can have in the same namespace workloads that are deployed using traditional deployment and others using serverless fashion. That will give us all the posibilities to design your cluster as you wish. To do that we will append the argument labels in a key=value fashion.
And we will get an output similar to this:
[ℹ] creating Fargate profile “fp-serverless-profile” on EKS cluster “cluster-test-hybrid”
[ℹ] created Fargate profile “fp-serverless-profile” on EKS cluster “cluster-test-hybrid”
If now we check the number of profiles that we have available we should get two profiles handling three namespaces (the ones that are managed by the default profile — default and kube-system — and the one — serverless — handled by the one we just created now)
We just will use the following command to delete the default profile:
And the output of that command should be similar to this one:
[ℹ] deleted Fargate profile “fp-default” on EKS cluster “cluster-test-hybrid”
And after that, we have now ready our cluster with limited scope for serverless deployments. In the next post of the series, we will just deploy workloads on both fashions to see the difference between them. So, don’t miss the updates regarding this series making sure that you follow my posts, and if you’d like the article, or you have some doubts or comments, please leave your feedback using the comments below!
EKS Fargate AWS Kubernetes Cluster: Learn how to create a Kubernetes cluster that can use also all the power of serverless computing using AWS Fargate
We know that there are several movements and paradigms that are pushing us hard to change our architectures trying to leverage much more managed services and taking care of the operational level so we can focus on what’s really important for our own business: creating applications and deliver value through them.
AWS from Amazon has been a critical partner during that journey, especially in the container world. With the release of EKS some time ago were able to provide a managed Kubernetes service that everyone can use, but also introducing the CaaS solution Fargate also gives us the power to run a container workload in a serverless fashion, without needing to worry about anything else.
But you could be thinking about if those services can work together. And the short answer is yes. But even more important than that we’re seeing that also they can work in a mixed mode:
So you can have an EKS cluster that has some nodes that are Fargate services and some nodes that are normal EC2 machines for workloads that are working in a state-full fashion or fits better in a traditional EC2 approach. And everything works by the same rules and is managed by the same EKS Cluster.
So, that sounds amazing but, How we can do that? Let’s see.
eksctl
To get to that point there is a tool that we need to introduce first, and that tool is named eksctl and it is a command-line utility that helps us to do any action to interact with the EKS service and simplifies a lot the work to do and also be able to automate most of the tasks in a non-human required mode. So, the first thing we need to is to get eksctl in our platforms ready. Let’s see how we can get that.
We have here all the detailed from Amazon itself about how to install eksctl in different platforms, no matter if you’re using Windows, Linux, or MacOS X:
Installing or updating eksctl – Amazon EKS
Learn how to install or update the eksctl command line tool. This tool is used to create and work with an Amazon EKS cluster.
After doing that we can check that we have installed the eksctl software running the command:
eksctl version
And we should get an output similar to this one:
eksctl version output command
So after doing that we can see that we have access to all the power behind the EKS service just typing these simple commands into our console window.
Creating the EKS Hybrid Cluster
Now, we’re going to create a mixed environment with some EC2 machines and enable the Fargate support for EKS. To do that, we will start with the following command:
eksctl create cluster --version=1.15 --name=cluster-test-hybrid --region=eu-west-1 --max-pods-per-node=1000 --fargate
[ℹ] eksctl version 0.26.0
[ℹ] using region eu-west-1
[ℹ] setting availability zones to [eu-west-1c eu-west-1a eu-west-1b]
[ℹ] subnets for eu-west-1c - public:192.168.0.0/19 private:192.168.96.0/19
[ℹ] subnets for eu-west-1a - public:192.168.32.0/19 private:192.168.128.0/19
[ℹ] subnets for eu-west-1b - public:192.168.64.0/19 private:192.168.160.0/19
[ℹ] using Kubernetes version 1.15
[ℹ] creating EKS cluster "cluster-test-hybrid" in "eu-west-1" region with Fargate profile
[ℹ] if you encounter any issues, check CloudFormation console or try 'eksctl utils describe-stacks --region=eu-west-1 --cluster=cluster-test-hybrid'
[ℹ] CloudWatch logging will not be enabled for cluster "cluster-test-hybrid" in "eu-west-1"
[ℹ] you can enable it with 'eksctl utils update-cluster-logging --region=eu-west-1 --cluster=cluster-test-hybrid'
[ℹ] Kubernetes API endpoint access will use default of {publicAccess=true, privateAccess=false} for cluster "cluster-test-hybrid" in "eu-west-1"
[ℹ] 2 sequential tasks: { create cluster control plane "cluster-test-hybrid", create fargate profiles }
[ℹ] building cluster stack "eksctl-cluster-test-hybrid-cluster"
[ℹ] deploying stack "eksctl-cluster-test-hybrid-cluster"
[ℹ] creating Fargate profile "fp-default" on EKS cluster "cluster-test-hybrid"
[ℹ] created Fargate profile "fp-default" on EKS cluster "cluster-test-hybrid"
[ℹ] "coredns" is now schedulable onto Fargate
[ℹ] "coredns" is now scheduled onto Fargate
[ℹ] "coredns" pods are now scheduled onto Fargate
[ℹ] waiting for the control plane availability...
[✔] saved kubeconfig as "C:\\Users\\avazquez/.kube/config"
[ℹ] no tasks
[✔] all EKS cluster resources for "cluster-test-hybrid" have been created
[ℹ] kubectl command should work with "C:\\Users\\avazquez/.kube/config", try 'kubectl get nodes'
[✔] EKS cluster "cluster-test-hybrid" in "eu-west-1" region is ready
This command will setup the EKS cluster enabling the Fargate support.
NOTE: The first thing that we should notice is that the Fargate support for EKS is not yet available in all the AWS regions. So, depending on the region that you’re using you could get an error. At this moment this is just enabled in US East (N. Virginia), US East (Ohio), US West (Oregon), Europe (Ireland), Europe (Frankfurt), Asia Pacific (Singapore), Asia Pacific (Sydney), Asia Pacific (Tokyo) based on the information from AWS Announcements: https://aws.amazon.com/about-aws/whats-new/2020/04/eks-adds-fargate-support-in-frankfurt-oregon-singapore-and-sydney-aws-regions/
So, now, we should add to that cluster a Node Group. a Node Group is a set of EC2 instances that are going to be managed as part of it. And to do that we will use the following command:
eksctl create nodegroup --cluster cluster-test-hybrid --managed
[ℹ] eksctl version 0.26.0
[ℹ] using region eu-west-1
[ℹ] will use version 1.15 for new nodegroup(s) based on control plane version
[ℹ] nodegroup "ng-1262d9c0" present in the given config, but missing in the cluster
[ℹ] 1 nodegroup (ng-1262d9c0) was included (based on the include/exclude rules)
[ℹ] will create a CloudFormation stack for each of 1 managed nodegroups in cluster "cluster-test-hybrid"
[ℹ] 2 sequential tasks: { fix cluster compatibility, 1 task: { 1 task: { create managed nodegroup "ng-1262d9c0" } } }
[ℹ] checking cluster stack for missing resources
[ℹ] cluster stack has all required resources
[ℹ] building managed nodegroup stack "eksctl-cluster-test-hybrid-nodegroup-ng-1262d9c0"
[ℹ] deploying stack "eksctl-cluster-test-hybrid-nodegroup-ng-1262d9c0"
[ℹ] no tasks
[✔] created 0 nodegroup(s) in cluster "cluster-test-hybrid"
[ℹ] nodegroup "ng-1262d9c0" has 2 node(s)
[ℹ] node "ip-192-168-69-215.eu-west-1.compute.internal" is ready
[ℹ] node "ip-192-168-9-111.eu-west-1.compute.internal" is ready
[ℹ] waiting for at least 2 node(s) to become ready in "ng-1262d9c0"
[ℹ] nodegroup "ng-1262d9c0" has 2 node(s)
[ℹ] node "ip-192-168-69-215.eu-west-1.compute.internal" is ready
[ℹ] node "ip-192-168-9-111.eu-west-1.compute.internal" is ready
[✔] created 1 managed nodegroup(s) in cluster "cluster-test-hybrid"
[ℹ] checking security group configuration for all nodegroups
[ℹ] all nodegroups have up-to-date configuration
So now we should be able to use kubectl to manage this new cluster. If you don’t have installed kubectl or you haven’t heard about it. This is the command line tool that allow us to manage your Kubernetes Cluster and you can install it based on the documentation shown here:
Installing or updating eksctl – Amazon EKS
Learn how to install or update the eksctl command line tool. This tool is used to create and work with an Amazon EKS cluster.
So, now, we should start taking a look at the infrastructure that we have. So if we type the following command to see the nodes at our disposal:
kubectl get nodes
We see an output similar to this:
NAME STATUS ROLES AGE VERSION
fargate-ip-192-168-102-22.eu-west-1.compute.internal Ready <none> 10m v1.15.10-eks-094994
fargate-ip-192-168-112-125.eu-west-1.compute.internal Ready <none> 10m v1.15.10-eks-094994
ip-192-168-69-215.eu-west-1.compute.internal Ready <none> 85s v1.15.11-eks-bf8eea
ip-192-168-9-111.eu-west-1.compute.internal Ready <none> 87s v1.15.11-eks-bf8eea
As you can see we have 4 “nodes” two that start with the fargate name that are fargate nodes and two that just start with ip-… and those are the traditional EC2 instances. And after that moment that’s it, we have our mixed environment ready to use.
We can check the same cluster using the AWS EKS page to see that configuration more in detail. If we enter in the EKS page for this cluster we see in the Compute tab the following information:
We see under Node Groups the data around the EC2 machines that are managed as part of this cluster and as you can see we saw 2 as the Desired Capacity and that’s why we have 2 EC2 instances in our cluster. And regarding the Fargate profile, we see the namespaces set to default and kube-system and that means that all the deployments to those namespaces are going to be deployed using Fargate Tasks.
Summary
In the following articles in these series, we will see how to progress on our Hybrid cluster, deploy workloads scale it based on the demand that we’re getting, enabling integration with other services like AWS CloudWatch, and so on. So, stay tuned, and don’t forget to follow my articles to not miss any new updates as soon as it’s available to you!
Managed Container Platform is disrupting everything. We’re living in a time where development and the IT landscape are changing, new paradigms like microservices and containers seem to be out there for the last few years, and if we trust the reality that the blog posts and the articles show today, we’re all of the users already using them all the time.
Did you see any blog posts about how to develop a J2EE application running on your Tomcat server on-prem? Probably not. The most similar article should probably be how to containerize your tomcat-based application.
But do you know what? Most companies still are working that way. So even if all companies have a new digital approach in some departments, they also have other ones being more traditional.
So, that seems that we need to find a different way to translate the main advantages of a container-based platform to a kind of speech they can see and realize the tangible benefits they can get from there and have the “Hey, this can work for me!” kind of spirit.
1. You will get all components isolated and updated more quickly
That’s one of the great things about container-based platforms compared with previous approaches like application server-based platforms. When you have an application server cluster, you still have one cluster with several applications. So you usually do some isolation, keep related applications, provide independent infrastructure for the critical ones, and so on.
But even with that, at some level, the application continues to be coupled, so some issues with some applications could bring down another one that was not expected for business reasons.
With a container-based platform, you’re getting each application in its bubble, so any issue or error will affect that application and nothing more. Platform stability is a priority for all companies and all departments inside them. Just ask yourself: Do you want to end with those “domino’s chains” of failure? How much will your operations improve? How much will your happiness increase?
Additionally, based on the container approach, you will get smaller components. Each of them will do a single task providing a single capability to your business, which means that it will be much easier to update, test, and deploy in production. So that, in the end will generate more deployments into the production environment and reduce the time to market for your business capabilities.
You will be able to deploy faster and have more stable operations simultaneously.
2.- You will optimize the use of your infrastructure
Costs, everything is about costs. There are no single conversations with customers who are not trying to pay less for their infrastructure. So, let’s face it. We should be able to run operations in an optimized way. So, if our infrastructure cost is going higher, that needs to mean that our business increases.
Container-based platforms will allow optimizing infrastructure in two different ways. First, if using two main concepts: Elasticity and Infrastructure Sharing.
Elasticity is related because I’m only going to have the infrastructure I need to support the load I have at this moment. So, if the load increases, my infrastructure will increase to handle it, but after that moment goes away, it will go back to what it needs now after that peak happened.
Infrastructure sharing is about using each server’s part that is free to deploy other applications. Imagine a traditional approach where I have two servers for my set of applications. Probably I don’t have 100% usage of those servers because I need to have some spare computer to be able to act when the load increases. I probably have 60–70% percent. That means 30% free. If we have different departments with different systems, and each has its infrastructure 30% free, how much of our infrastructure are we just throwing away? How many dollars/euros/pounds are you just throwing off the window?
Container-based platforms don’t need specific tools or software installed on the platform to run a different kind of application. It is not required because everything resides inside the container, so I can use any free space to deploy other applications doing a more efficient usage of those.
3.- You will not need infrastructure for administration
Each system that is big enough has some resources dedicated to being able to manage it. However, even most of the recommended architectures recommend placing those components isolated from your runtime components to avoid any issue regarding administrator or maintenance that can affect your runtime workloads, which means specific infrastructure that you’re using for something that isn’t helping your business. Of course, you can explain to any business user that you need a machine to run that provides the capabilities required. But it is more complex than using additional infrastructure (and generating cost) to place other components that are not helping the business.
So, managed container platforms take that problem away because you’re going to provide the infrastructure you need to run your workloads, and you’re going to be given for free or such low fee the administration capabilities. And addition to that, you don’t even need to worry that administration features are always available and working fine because this is leverage to the provider itself.
Wrap up and next steps
As you can see, we describe very tangible benefits that are not industry-based or development focus. Of course, we can have so many more to add to this list, but these are the critical ones that affect any company in any industry worldwide. So, please, take your time to think about how these capabilities can help to improve your business. But not only that, take your time to quantify how that will enhance your business. How much can you save? How much can you get from this approach?
And when you have in front of you a solid business case based on this approach, you will get all the support and courage you need to move forward during that route!! So I wish you a peaceful transition!
Find a way to re-define and re-organize the name of your Prometheus metrics to meet your requirements
Prometheus has become the new standard when we’re talking about monitoring our new modern application architecture, and we need to make sure we know all about its options to make sure we can get the best out of it. I’ve been using it for some time until I realized about a feature that I was desperate to know how to do, but I couldn’t find anywhere clearly define. So as I didn’t found it easily, I thought about writing a small article to show you how to do it without needed to spend the same time as I did.
We have plenty of information about how to configure Prometheus and use some of the usual configuration plugins, as we can see on its official webpage [1]. Even I already write about some configuration and using it for several purposes, as you can see also in other posts [2][3][4].
One of these configuration plugins is about relabeling, and this is a great thing. We have that each of the exporters can have its labels and meaning for those, and when you try to manage different technologies or components makes complex that all of them match together even if all of them follow the naming convention that Prometheus has [5].
But I had this situation, and I’m sure you have gone or will go towards that as well, that I have similar metrics for different technologies that for me are the same, and I need to keep them with the same name, but as they belong to other technologies they are not. So I need to find a way to rename the metric, and the great thing is that you can do that.
To do that, you just need to do a metric_relabel configuration. This configuration applies to relabel (as the name already indicates) labels of your prometheus metrics in this case before being ingested but also allow us to use some notable terms to do different things, and one of these notable terms is __name__. __name__ is a particular label that will enable you to rename your prometheus metrics before being ingested in the Prometheus Timeseries Database. And after that point, this will be as it will have that name since the beginning.
How to use that is relatively easy, is as any other relabel process, and I’d like to show you a sample about how to do it.
- source_labels: [__name__]
regex: 'jvm_threads_current'
target_label: __name__
replacement: 'process_thread_count'
Here it is a simple sample to show how we can rename a metric name jvm_threads_current to count the threads inside the JVM machine to do it more generic to be able to include the threads for the process in a process_thread_count prometheus metrics that we can use now as it was the original name.
Add a header to begin generating the table of contents
We all know that in the rise of the cloud-native development and architectures, we’ve seen Kubernetes based platforms as the new standard all focusing on new developments following the new paradigms and best practices: Microservices, Event-Driven Architectures new shiny protocols like GraphQL or gRPC, and so on and so forth.
This article is part of my comprehensive TIBCO Integration Platform Guide where you can find more patterns and best practices for TIBCO integration platforms.
