DevOps and Docker Talk: Cloud Native Interviews and Tooling

Better K8s Prometheus Alerts with Robusta

Mar 4, 2023

Natan Yellin, Co-founder of Robusta.dev, discusses enhancing Kubernetes alerts with Robusta, improving Prometheus messages, deploying Robusta in clusters, using push notifications, and dealing with CPU limits in Kubernetes. The podcast covers challenges in alert management, monitoring solutions with Prometheus, effective alert management, push notifications for alerting, and detailed explanations of CPU limits.

Ask episode

Chapters

Transcript

Episode notes

Intro

00:00 • 4min

Challenges in Managing Alerts in Kubernetes Environments and the Role of Robusta

04:14 • 10min

Exploring Monitoring Challenges and Solutions with Prometheus in Kubernetes Environment

13:50 • 5min

Effective Management of Prometheus Alerts and Monitoring with Robusta

18:39 • 20min

Exploring Push Notifications for Alert Management

38:15 • 2min

Tool for setting up alerts and detailed explanation of CPU limits in Kubernetes

39:52 • 10min

Bret is joined by Natan Yellin, the co-founder of Robusta.dev to talk Kubernetes and Prometheus monitoring, alerting, and maybe some CPU limit ranting.

Robusta tries to fill the gap left by Kubernetes' own AlertManager which has a very specific and not so helpful way of describing events in your cluster. This makes it hard to diagnose the cause of the event and you're left with Google, StackOverflow and an awful lot of head-scratching. Robusta acts as a proxy between AlertManager and your notification platform of choice.

In the show we talk about what Robusta is, how to deploy it in your clusters, and Natan also details some of the enhancements in their cloud offering that you can layer on top of that, which has a generous free tier.

Streamed live on YouTube on January 5, 2023.

Unedited live recording of this show on YouTube (Ep. #197). Includes demos.

★Topics★
Robusta Website
Robusta on GitHub
KubeCon - Building a Runbook Automation System for Prometheus and Kubernetes
Stop using K8s CPU limits
Recommended Pod Spec
Send Push notifications to your phone
Prometheus AlertManager
Grafana Labs
Kubewatch

★Natan Yellin★
Natan on Twitter
Natan on LinkedIn

★Join my Community★
New live course on CI automation and gitops deployments
Best coupons for my Docker and Kubernetes courses
Chat with us and fellow students on our Discord Server DevOps Fans
Grab some merch at Bret's Loot Box

Homepage bretfisher.com

(00:00) - DDT MAIN
(00:04) - Intro
(00:53) - In today's episode
(02:59) - Main show
(03:27) - Introducing Natan
(03:53) - Alert fatigue
(04:29) - Where did the idea for Robusta come from?
(08:16) - Someone has to do the job
(09:17) - What does Robusta offer?
(10:25) - Proxying the alerts and providing context
(11:30) - Saving 10 to 30 minutes
(13:48) - The open source Robusta repo
(14:10) - The need to de-aggregate event data
(15:09) - Example or demo
(15:39) - Question about observability for microservices
(18:38) - Tip 1 Consider using silences
(19:49) - Tip 2 Monitor outcomes
(20:23) - Don't ignore alerts because of fatigue
(23:13) - Sending to different channels based on priority
(24:42) - Question about sending messages to destinations
(26:17) - Question
(26:49) - Installing Robusta
(27:42) - Demo set up commands
(27:54) - Questions
(28:11) - Demo Kubernetes-specific
(29:05) - Multi-cluster question
(31:32) - What does the SaaS platform do?
(32:44) - Demo with SaaS
(33:37) - kubectl not recommended
(35:03) - Breaking the glass
(38:15) - Question about notifications
(40:14) - Getting started
(41:24) - CPU limiting
(42:15) - Soft limits on CPU in Kubernetes
(44:35) - Bret's pod spec
(49:22) - Outro

You can also support my content by subscribing to my YouTube channel and my weekly newsletter at bret.news!

Grab the best coupons for my Docker and Kubernetes courses.
Join my cloud native DevOps community on Discord.
Grab some merch at Bret's Loot Box
Homepage bretfisher.com