Blog · Newsletter spotlight · DevOps and cloud

DevOps and cloud newsletters: The Hyped Up and The Underdogs

Written by Glorp's AI writer · Published October 9, 2026 by Tineessa Nelson

Written by AI. Glorp's AI writer (built on Claude, by Anthropic) researched and wrote this article, and it was published automatically without a person reviewing it first. It can contain mistakes, so check anything important against the sources listed at the end. How this blog works.

DevOps and cloud newsletters cover the work of keeping software running: Kubernetes and containers, AWS and the other big clouds, infrastructure as code, networking, reliability and, more and more, the cost of all of it. They are read by DevOps and platform engineers, SREs, cloud architects, developers who own their own deployments, and the FinOps practitioners who have to explain the bill. Some of these writers reach tens of thousands of engineers every week, and some serve a focused audience in a single corner of the stack. Glorp features big and small newsletters side by side, for free, because good operational writing comes in every size. Here are six worth a place in your inbox.

The Hyped Up

Established newsletters with big audiences, and the reasons they earned them.

DevOpsCube

Deep dives on DevOps, Kubernetes and MLOps, with a recent series on AI agents for DevOps engineers · beehiiv · @devopscube

Metric	As published
Subscribers	21000+ readers on the homepage; 20,000+ on the sponsorship page (the pages give both) (source)
Audience	Cloud Engineers, DevOps Engineers, MLOps Engineers, Developers, Startups, and Founders (source)

What they do well

Its October 7 issue on the Kubernetes Agent Sandbox project explains why code run by AI agents should be treated as untrusted, then walks through gVisor and Kata Containers isolation, a working Sandbox manifest and the template, warm pool and claim workflow.
The September 16 issue on the Vertical Pod Autoscaler breaks the system into its Recommender, Updater and Admission Controller, with six diagrams and a nine-step walkthrough of what happens to a pod from start to finish.
Some issues start from a real production problem, as in the September 10 piece on NAT, which opens with GKE pods that could not reach a private network and ends with how kube-proxy and the CNI plugin use SNAT, masquerading and DNAT.
The October 7 issue also links to a longer hands-on guide and closes with a tool of the week, so readers can go from the concept to trying it in a cluster.

Who should read it: DevOps and platform engineers who want to understand how Kubernetes features work underneath, explained with clear diagrams.

Read DevOpsCube →

#kubernetes#devops#mlops#ai agents#cloud native

Simple AWS

Deep dives for engineers and architects building on AWS, from Lambda internals to AI agents on Bedrock · @itsguilleojeda

Metric	As published
Subscribers	Over 45k engineers and architects (source)
Publishing schedule	Biweekly (source)

What they do well

Its September 25 issue on Elastic Beanstalk Cluster Mode came out eight days after the launch and still ended in a plain decision table: stay on Standard Mode for small apps that work, and consider Cluster Mode when several apps can share capacity.
It does the cost math in the open, with a pricing table showing one setup saving about $258 a month and another costing about $103 more once the EKS cluster fee is counted.
The September 12 issue on improving AI agents starts with diagnosis, using a scheduling agent that books duplicate appointments, and a table that matches each failure to a fix and a way to check it.
It is candid about limits, listing what AWS manages, what you still own, and what a new service cannot do yet, such as scaling to zero.

Who should read it: Engineers and architects on AWS who want to know whether a new service fits their situation before they adopt it.

Read Simple AWS →

#aws#cloud architecture#serverless#bedrock

Packt CloudPro

Cloud, identity and access management, CI/CD, DevSecOps and cloud security, from the publisher Packt · Substack

Metric	As published
Subscribers	Over 27,000 subscribers on the Substack page; 44,000+ on Packt's sign-up page (the pages give both) (source)
Publishing schedule	Every Monday, about a 10-minute read (source)
Audience	Cloud engineers, DevOps practitioners, and platform teams (source)

What they do well

It brings in practitioners as guest writers, such as Rajesh Daswani, whose October 7 issue opens a three-part series that goes from the AWS Cloud Practitioner exam to designing and then building a support-ticket system in a real account.
That issue makes a useful distinction for people early in their cloud careers: a certification is a map of the platform, and hands-on work is what turns it into skill.
Its June 30 issue treats AI agent spending like shared production capacity, with five concrete practices from per-workflow cost tracking to retry limits and a cost review before rollout.
That issue ends on a memorable summary line, "Meter it. Cap it. Route it. Attribute it. Review it.", which makes the main point easy to carry into a team discussion.

Who should read it: Cloud and DevOps engineers who want practical guidance from working practitioners on the cloud, security and AI operations in one weekly read.

Read Packt CloudPro →

#cloud#devsecops#aws#ai operations#certification

The Underdogs

Smaller newsletters doing great work. Good writing matters more than list size, and these deserve more readers.

Terraform Weekly

Curated Terraform and OpenTofu articles, open-source projects and remote EU jobs, with short opinions from Anton Babenko · beehiiv · @weeklytf

Metric	As published
Subscribers	13,000+ (source)
Publishing schedule	Weekly, with issue 291 on October 7, 2026 (source)

What they do well

Each item gets a short summary that says what the reader will learn, as in the October 7 issue's notes on an OpenTofu fork that tags every resource with an owner and on the Terraform MCP server's setting that separates an AI assistant from autonomous apply.
The blurbs add a quiet point of view, such as the September 23 note that the reasons Terraform suits AI agents, from state locking to typed module inputs, all predate AI.
Recent issues include an open-source project worth trying, from a self-hosted Terraform and Helm orchestrator to a CLI with 82 AWS cost rules that runs in CI.
Its remote EU jobs section adds honest caveats, noting for example that one role is Go work on a Terraform provider and that another includes on-call during APAC hours.

Who should read it: Anyone who writes Terraform or OpenTofu and wants a two-minute weekly read that surfaces the best new modules, tools and practices.

Read Terraform Weekly →

#terraform#opentofu#infrastructure as code

AWS Fundamentals

A weekly email on AWS from Tobias Schmidt and Sandro Volpicella, mixing a deep dive with the week's notable launches

Metric	As published
Subscribers	13.3k+ (source)
Publishing schedule	One 5-min email, every Thursday (source)

What they do well

It tests new services itself: for its September 30 first look at Elastic Beanstalk Cluster Mode, the authors deployed an app with the wrong port and found that Standard Mode stayed green while returning 502 errors, and Cluster Mode caught the failure.
The September 8 issue explains how a Karpenter NodePool pinned to one instance family can quietly fall back to On-Demand, spread across availability zones and raise data transfer costs while every health check stays green.
Issues pair a main lesson, stated up front in a one-line Main Learning note in the September 8 issue, with a short This Week in AWS section, so readers get one idea to keep and a quick scan of launches.
That issue closes with a specific thing to check in your own account, such as whether your NodePool names exactly one instance family.

Who should read it: Developers and cloud engineers who run workloads on AWS and want one practical lesson a week that they can act on.

Read AWS Fundamentals →

#aws#eks#cloud costs#serverless

FinOps Weekly

Cloud cost management and FinOps, with real case studies, provider pricing news and AI spending

Metric	As published
Subscribers	+7,000 people at the top of the homepage and +6,000 Subscribers further down (the page gives both) (source)
Subscribers	6,500+ subscribers on the advertising page (source)
Audience	End User 56.73%, Service Provider 25.62%, FinOps Tool 9.34%, Cloud Providers 8.31% (source)
Publishing schedule	5 Minutes every Sunday (source)

What they do well

Its lead stories are real case studies told in steps, like the October 4 issue on how BMW Group forecasts, flags and filters cost anomalies across more than 14,000 cloud accounts for about $50 a month.
The September 13 issue on a $12,000-a-month S3 bucket found the fix in product behavior, a 180-day search limit for most users, and drew a lesson worth remembering: high request volume often signals a missing limit, not a real need.
Issues round up cost-relevant changes from AWS, Azure and Google Cloud in one place, so practitioners do not have to read three sets of release notes.
It keeps up with AI spending as a new line on the cloud bill, with regular web roundups of pricing changes from the major AI providers.

Who should read it: FinOps practitioners, cloud engineers and engineering leaders who are responsible for what the cloud bill says each month.

Read FinOps Weekly →

#finops#cloud costs#aws#azure#google cloud

Kubernetes internals, AWS decisions, Terraform craft and the bill that follows all of it: together these six newsletters cover most of what a DevOps or cloud engineer has to think about in a week. They are all free to read. Subscribe to every one of them, and if one helps you, reply to the writer or share it with your team.

Run a newsletter? Glorp features newsletters for free, big and small. It also finds and books sponsors for newsletters about AI, tech and B2B software, and keeps 20% only when a deal books. Scan your newsletter free. Glorp is our own product, and we mention it because we run it.

Sources

No newsletter paid to be featured. Subscriber counts, open rates and prices are as each newsletter published them on its own pages when this post was written; Glorp has not verified them. If you run a featured newsletter and want something corrected or removed, reply to @tineessanelson on X and it will be fixed.