DevOps & Infrastructure

Moldova-based hosting chosen for a different jurisdiction and looser content policy — which also raises the question of who your neighbours are.

The cloud with the widest service range. Compared with Azure and Google Cloud the breadth is an advantage, but cost tracking is its own discipline — egress charges in particular catch people out.

Configuration management and automation that provisions servers with YAML playbooks. Unlike Puppet and Chef, it needs no agent on the target and works over SSH.

An orchestrator scheduling data pipelines as Python-defined directed graphs (DAGs). Unlike cron it offers dependencies, retries and visibility into historical runs.

A module architecture plus per-directory configuration through .htaccess. Still dominant in shared hosting, and hungrier for memory than nginx under high concurrency.

A suite of Kubernetes-native workflow and GitOps continuous-delivery tools (Argo CD, Workflows). Unlike CI servers such as Jenkins, it runs inside the cluster and treats Git as the single source of truth, continuously syncing cluster state to the repo.

A VPS and dedicated server provider. Compared with the big clouds it is budget-focused with a narrow range of managed services — for small-scale projects.

A list of ready actions for GitHub Actions, automating your workflow. Unlike the Marketplace it is curated, grouping proven actions by use.

A list of roles and resources for Ansible, the Python-based IT configuration tool. Unlike Puppet and Chef, it works over SSH with no agent installed on the target machine.

A resource list for Backstage, building internal developer portals. Unlike general DevOps lists, it gathers plugins, deployment examples and write-ups for this Spotify-born portal.

A resource list for the AWS CDK, defining infrastructure with real code. Unlike Terraform, you write infrastructure in general-purpose languages such as TypeScript or Python instead of HCL, and it compiles to CloudFormation.

A list of tools, tutorials and production experience for the Kubernetes ecosystem. Unlike the official docs it gathers community tools and real-world patterns together.

A list of tools, libraries and guides for the Nix package manager and NixOS. Unlike general Linux lists it focuses on reproducible, declarative configuration.

A resource list for OpenTofu, the open-source fork of Terraform. Unlike Terraform, it stays under an open-source (MPL) licence at the Linux Foundation; it was born when HashiCorp moved to the BSL in 2023.

A list of exporters and resources for Prometheus, the open-source monitoring system. Unlike push-based monitoring it pulls metrics from targets; the list gathers exporters, dashboards and tools.

A resource list for Splunk, the log-search and monitoring platform. Unlike lists for open-source stacks such as ELK or Loki, it focuses on the commercial Splunk ecosystem — apps, SPL queries, add-ons.

A list of modules and resources for Terraform, managing infrastructure as code. Unlike OpenTofu lists, it centres on HashiCorp’s original tool and its older, broader module ecosystem.

The standard for background tasks in Python, using Redis or RabbitMQ as a broker. Compared with simpler queues such as RQ it is comprehensive — scheduled tasks (beat), retries, workflow primitives — at the cost of heavier configuration.

The EFF client that obtains and renews Let's Encrypt certificates; the site gives exact commands once you pick your web server and OS. Unlike lightweight clients such as acme.sh, it can edit nginx and Apache configuration itself.

A cloud-native, Kubernetes-integrated chaos-engineering (fault-injection) platform. Unlike Chaos Monkey, which only kills instances, you declare faults such as network latency, disk errors and clock skew as Kubernetes resources.

Infrastructure offering CDN, DNS, DDoS protection and edge computing in one platform. Unlike separate services it unites them in one dashboard with a generous free tier.

Exposes a local web server to the internet over a temporary trycloudflare.com HTTPS URL with a single cloudflared command and no account; avoids the signup and interstitial limits of free ngrok.

Kubernetes certifications through CKA, CKAD and CKS. Unlike cloud provider certificates they are vendor-neutral, and the exam happens at a terminal — not multiple choice, but tasks solved on a live cluster.

Runs development environments as containers on your own infrastructure, reached from a VS Code or JetBrains client. Against GitHub Codespaces it is not tied to a cloud provider — both the compute and the code stay with you.

A self-hosted web interface for cron jobs that renders expressions in readable form and keeps a run log. Against crontab -e it shows the next run time and the last output, which makes silently failing jobs visible.

Builds a cron expression from a form and states in words when it will fire. Against crontab.guru it runs the other way: instead of decoding an expression, it produces one from the schedule you describe.

Runs an Android emulator in a container with VNC access to the screen. Compared with cloud device farms it is free and runs in your own CI — an emulator for automated tests without standing up devices.

Brings a Vercel-like deploy experience to your own VPS — git-push deploys, database provisioning, automatic TLS. It fills the hole left when Heroku’s free tier closed.

A build tool with a syntax blending Dockerfile and Makefile, giving the same result locally and in CI. Unlike a plain Makefile every step is isolated in a container; the project is no longer actively maintained, so factor that into new setups.

Brendan Gregg's page on the flame graph visualisation: how to read one, the CPU/memory/off-CPU variants, generation tools and talks. You learn from its creator how to show profiler output as a stack tree instead of a flat list.

A local AWS emulator for development and CI that needs no account or token; the AWS SDK, CLI and Terraform simply point at port 4566. Its difference from LocalStack is that it aims to be a drop-in image swap without feature gates or authentication.

Analyses a Dockerfile, explains what happens layer by layer and suggests improvements. Against hadolint it is visual and states why each rule exists, which makes it usable for learning to write Dockerfiles rather than only linting them.

Renders repository history as a growing animated tree where you see who touched what. Not an analysis tool — a way to narrate a project’s history.

Unifies Prometheus, Loki, PostgreSQL and dozens of sources on one dashboard. Unlike dashboards tied to one tool it is data-source agnostic, which is how it became the shared face of monitoring stacks.

Turkey-based hosting and domains with .tr registration, local payment and Turkish support. Unlike foreign providers the server sits in Turkey — practical when data must stay in the country.

A private container-image registry with security scanning and access control, a CNCF graduated project. Unlike Docker Hub, it runs on your own infrastructure and scans images for vulnerabilities before release.

A package manager that packages and deploys Kubernetes applications as charts. Unlike plain YAML manifests, it adds templating, values files for environment differences and release rollback.

Dedicated servers and VPS in Germany and Finland. Compared with the big clouds the same cores and RAM cost a fraction; there are no managed services, so operations are entirely yours.

Hands-on labs and challenges teaching Linux, containers, Kubernetes and networking in real VMs inside the browser. Unlike articles that only explain, each lesson comes with a running environment.

Makes GitHub repos that don’t publish jars usable directly as dependencies. Unlike publishing to Maven Central, no registration step is needed — it builds a tag and serves it as a package.

Generates Kubernetes manifests from a form, explaining each field beside it. Against copying from the docs you leave with a valid manifest and pick up what the fields mean on the way.

A fast, interactive terminal UI for managing Kubernetes clusters. Instead of typing kubectl commands one by one, you move through resources by keyboard and jump straight into logs and shells.

A workflow orchestrator whose flows are declared in YAML. Unlike Airflow it uses YAML instead of Python, so it works for teams not on Python and sits well in version control.

The official site of the open-source system that deploys, manages and heals containers at scale. Unlike Docker Compose it scales across machines and self-heals; in exchange the learning and operating costs are high.

A checklist to review before running Kubernetes in production: resource limits, health checks, networking, security and observability. Unlike general tutorials it is framed as 'check this before going live'.

Kelsey Hightower's guide to bootstrapping a Kubernetes cluster by hand, component by component, without automation. It exposes the certificates, etcd and API server details that installers hide, making it the standard route to understanding cluster internals. It differs from installers by showing the steps they hide.

A commercial server that reads Apache configuration as-is while using an event-driven core and fewer resources. Common in WordPress hosting alongside its cache plugin.

The LiteSpeed web server’s configuration reference, including how LSCache rules are written. Unlike Apache it is event-driven yet still reads .htaccess; cache settings account for most of the performance you get.

Scenarios carry branching, loops and error handling, which makes complex flows easier than Zapier’s linear steps. The learning curve is correspondingly steeper.

Workflow automation with a visual node editor. Unlike Zapier, executions are unmetered when self-hosted and you can drop JavaScript into a node, so you do not hit the no-code wall.

Node reference, self-hosting and credential management. Unlike community tutorials it covers production topics too — the queue-mode scaling section matters for anyone running it in production.

Runs a browser in a container and shares it over WebRTC, with several people able to control one session. Unlike remote-desktop tools it shares only the browser — for watching together and isolated browsing.

Its event-driven architecture carries tens of thousands of concurrent connections on little memory. It is used as a reverse proxy and load balancer more often than as a web server.

Registers the domain in its own name on your behalf, so you do not appear in WHOIS. Unlike a privacy proxy the legal ownership sits with them, which makes it a question of trust.

A community-driven open-source fork of Terraform that runs the same configurations. Unlike Terraform it stays under an open (MPL) licence at the Linux Foundation and adds its own features such as state encryption.

A build system that scales to large, multi-language monorepos (originated at Twitter). Compared with Bazel it is friendlier to the Python ecosystem and lighter to configure, since it infers dependencies from source.

An automation platform where any step can be Node.js or Python code pulling npm or PyPI packages. Unlike visual automation such as Zapier, it closes the gap to real code.

Production process manager for Node.js and scripts with built-in load balancing, crash restarts, and log streaming; offers zero-downtime reloads and simpler clustering than raw systemd units.

A self-hosted tool that manages Docker and Kubernetes environments through a visual UI. Instead of the command line you manage containers, stacks and logs from a panel — for watching several environments in one place.

A collection of real post-mortems of outages and system failures published by companies. Unlike general DevOps lists it focuses only on incident write-ups, useful for learning from failures and for interviews.

An automation tool with native SharePoint, Teams and Dataverse connectors. Compared with Zapier, the real reason it gets chosen is licensing — if the organisation is already on Microsoft 365 it is usually paid for.

A flow is an ordinary Python function you decorate into orchestration. Moving local code to production is shorter than learning Airflow’s separate DSL.

The Turkish translation of Git’s official book. Unlike command cheat sheets, its internals chapter — object model, refs — makes Git make sense instead of being memorised.

An open-source monitoring system for metrics collection and alerting, with a time-series database. Unlike push-based systems it pulls metrics from targets and queries them with PromQL.

The official six-step series on queue patterns — work queues, pub/sub, routing, RPC. Unlike Kafka, which is for event streaming, this is about task distribution.

The reference for its data structures — lists, sets, sorted sets, streams. Unlike Memcached it has those structures and persistence; treating Redis as a plain key-value store means leaving most of the tool unused.

Collects errors with stack traces, source maps and user context, grouping duplicates. Against reading log files, it tells you how many users a given error actually hit.

A static analyser for shell scripts that finds unquoted variables, wrong test syntax and portability traps. Against bash -n it checks meaning rather than syntax, so it catches the script that runs and misbehaves.

An education channel covering DevOps, Kubernetes, Docker, networking and cloud with visual explanations from beginner level. Unlike the official docs, it builds each concept in a diagram first and then moves to practice.

Records workflow state as an event history, so a crashed process resumes at the same point. Unlike task queues such as Celery the state is durable — built for multi-day operations that cannot simply be restarted.

HashiCorp’s tool to define and provision infrastructure as code, managing cloud resources declaratively. Unlike CloudFormation it is not tied to one cloud; it has been BSL-licensed since 2023 (IBM bought HashiCorp in 2025; the licence did not change), with OpenTofu as the openly licensed fork.

Frontend-focused deployment whose per-PR preview environments became an industry standard. Compared with Netlify, it is the company behind Next.js, so integration with that framework is tightest here.

A YAML validator that surfaces indentation and quoting errors and prints the parsed document back. Against a JSON validator it catches YAML’s actual trap: indentation that silently produces the wrong structure.

The automation tool with the widest integration catalogue — the niche SaaS you want is probably there. Compared with n8n the breadth is an advantage; pricing is per task, so it gets expensive quickly as volume grows.