---
title: "MRIDUL TIWARI (@MridulTi) | Hashnode"
description: "Learning in public. Read the latest articles by MRIDUL TIWARI on Hashnode."
url: "https://hashnode.com/@MridulTi"
---

@MridulTi

# MRIDUL TIWARI

@MridulTi · Joined March 2022

Learning in public

### [When /var/tmp forgot our Kafka directory](https://mriduliti.hashnode.dev/when-var-tmp-forgot-our-kafka-directory?utm_source=hashnode&utm_medium=feed)

The alert wasn’t dramatic. REST on 8083 answered fine, the connector showed RUNNING, and then I pulled task status and saw FAILED with a stack trace long enough to scroll past three screens. Patch nig

6 min read · [mriduliti.hashnode.dev](https://mriduliti.hashnode.dev/when-var-tmp-forgot-our-kafka-directory?utm_source=hashnode&utm_medium=feed)

[Oct 11](https://hashnode.com/posts/when-var-tmp-forgot-our-kafka-directory/6acb324db8a6d28dc5d7e82a)

### [When the cluster looked idle but nothing would schedule](https://mriduliti.hashnode.dev/when-the-cluster-looked-idle-but-nothing-would-schedule?utm_source=hashnode&utm_medium=feed)

The first thing I checked was the obvious one: CPU. Our non-prod EKS workers are Karpenter-managed, and application pods had been sitting in Pending long enough that Istio and app syncs were backing u

6 min read · [mriduliti.hashnode.dev](https://mriduliti.hashnode.dev/when-the-cluster-looked-idle-but-nothing-would-schedule?utm_source=hashnode&utm_medium=feed)

[Oct 4](https://hashnode.com/posts/when-the-cluster-looked-idle-but-nothing-would-schedule/6ac1c9fbe8a623698126213b)

### [Patch night wasn’t one incident — it was three failure modes in a row](https://mriduliti.hashnode.dev/patch-night-wasn-t-one-incident-it-was-three-failure-modes-in-a-row?utm_source=hashnode&utm_medium=feed)

Tuesday’s plan looked boring on paper: run the SSM patch runbooks across a bunch of Auto Scaling Groups, bake AMIs, promote launch templates, go home. By midnight I was still on a call with Jenkins co

9 min read · [mriduliti.hashnode.dev](https://mriduliti.hashnode.dev/patch-night-wasn-t-one-incident-it-was-three-failure-modes-in-a-row?utm_source=hashnode&utm_medium=feed)

[2](https://hashnode.com/posts/patch-night-wasn-t-one-incident-it-was-three-failure-modes-in-a-row/6ab889243239b46f2800a824) [Sep 27](https://hashnode.com/posts/patch-night-wasn-t-one-incident-it-was-three-failure-modes-in-a-row/6ab889243239b46f2800a824)

### [No official Storm Helm chart — how we put Apache Storm on EKS anyway](https://mriduliti.hashnode.dev/no-official-storm-helm-chart-how-we-put-apache-storm-on-eks-anyway?utm_source=hashnode&utm_medium=feed)

The Argo CD app for Storm had been red for two days, but the symptom that actually sent me down the rabbit hole was quieter than a crash loop: init containers on the Supervisor pods, stuck forever, pr

10 min read · [mriduliti.hashnode.dev](https://mriduliti.hashnode.dev/no-official-storm-helm-chart-how-we-put-apache-storm-on-eks-anyway?utm_source=hashnode&utm_medium=feed)

[Sep 17](https://hashnode.com/posts/no-official-storm-helm-chart-how-we-put-apache-storm-on-eks-anyway/6aac2b8dd9c9be5b83440eaa)

### [We tagged EKS workers like app servers — and Prometheus started scraping telegraf on nodes that never had it](https://mriduliti.hashnode.dev/we-tagged-eks-workers-like-app-servers-and-prometheus-started-scraping-telegraf-on-nodes-that-never-had-it?utm_source=hashnode&utm_medium=feed)

The Telegraf Down alerts started piling up on a Thursday morning, and at first glance it looked bad. Nine targets, all critical, all under the same rule name. My first instinct was a fleet-wide agent

7 min read · [mriduliti.hashnode.dev](https://mriduliti.hashnode.dev/we-tagged-eks-workers-like-app-servers-and-prometheus-started-scraping-telegraf-on-nodes-that-never-had-it?utm_source=hashnode&utm_medium=feed)

[Sep 12](https://hashnode.com/posts/we-tagged-eks-workers-like-app-servers-and-prometheus-started-scraping-telegraf-on-nodes-that-never-had-it/6aa4e191c617a90948586fd5)
