What’s New?
K8s CloudWatch & Prometheus Adapters Are Now Available
Typically, Kubernetes HPA (Horizontal Pod Autoscaler) scales your pods based on a very simple rule that you set up in your HPA manifest file. You can now scale your Kubernetes deployment using metrics from AWS CloudWatch using K8s CloudWatch Adapter.
For example, it can read the number of messages in your AWS SQS queue and scale up your deployment depending on a specific number-based rule that you pre-set. In this case, your cluster can handle the traffic more smoothly than before. Unlike traditional HPA rules which scale up only when it meets the physical resource limit, which can cause throttling, by using CloudWatch metrics it can scale up before it meets the physical limit using an outside cluster indicator.
https://github.com/awslabs/k8s-cloudwatch-adapter
The CloudWatch adapter is a useful tool if your deployment needs to scale based on some of your AWS resources. However, it cannot scale based on cluster metrics, such as the number of connections to the DB running in your Kubernetes cluster. Prometheus scrapes these kinds of metrics, and the adapter utilizes this in order to scale up your deployment.
EKS 1.18
Kubernetes version 1.18 supports Server-side Apply beta. This feature tracks and manages changes to fields of all K8s objects. This allows you to know what changed your resource and when. With a new pathType and IngressClass resource having been added to the ingress specification, these features make it simpler to customize ingress configuration. You can learn more about this version here.
Elasticsearch Dashboard
Elasticsearch logging allows you to deploy the AWS Elasticsearch Service dashboard. This dashboard visualizes search rate/latency, index rate/latency, cluster status, HTTP request response code, invalid host header request, free storage, max CPU utilization and more. This can provide insights to help you better understand the current state of your Elasticsearch cluster to conduct preventive maintenance. Want to learn more? Let us know.
Tales From the Ops Side
In case you missed it, Tales From the Ops Side’s newest episode, Dr. Strangecron or How I Learned to Stop Worrying and Love the Lastcomm, is now available to stream on your favourite podcast platform!
After having lunch interrupted by an urgent call from a client, stack.io CEO Hany Fahim put on his Sherlock Holmes hat to investigate a missing file full of sensitive data.
Episode 3 of Tales From the Ops Side is also released online, and takes you to Brazil for mystery and intrigue!
This story is about the ramifications of a political event thousands of miles away.
Hany was at home one evening in 2015 when he received an alert about a huge spike in traffic. Normally, this would point to a number of things such as a marketing event or even an attack—but nothing seemed to add up.
Hany traced the users’ IP addresses and discovered they all seemed to lead back to Brazil. Was this a calculated attack? He turned to the social media grapevine for answers.
Sure enough, Twitter revealed that the Brazilian government had placed a 48-hour ban on WhatsApp, a social networking app.
You can find out what happened next and how Hany dealt with the challenge by going to Tales From The Ops Side.
On this dedicated podcast website, you’ll also find episode 2. In this tale, Hany takes listeners on an incredible journey, exploring the connection between earthquakes, leap seconds, and the moon, and, closer to home, crashed websites. Discover Earthquakes and The Moon here.
You can subscribe to stack.io’s podcast show on your favourite podcast app, as well as our website.
If you like the show, please leave a review on iTunes!
Useful Reading
Manually create your own Kubernetes clusters anywhere you choose
You can now package and deploy Lambda functions as container images
Larger & faster io2 Block Express EBS Volumes with higher throughput
Amazon EBS gp3 volume lets you provision performance apart from capacity
Fun Stuff
Wish you could Skype someone in complete isolation? Run Skype inside an isolated Docker container on your Linux desktop!
Run NES emulator in your container.