{"canonical_url":"https://engineer.company/glossary/azure/","content_markdown":"# Azure\n\n## What it is\n\nAzure, or Microsoft Azure in full, is a catalogue of computing rented out from\nMicrosoft's own data centres: virtual machines, storage accounts, managed\ndatabases, container hosting, private networks and the tools to run them, paid\nfor by use. Microsoft sold it as Windows Azure until April 2014, when it took\nits present name.\n\nEvery request to manage Azure resources, whether a click in the portal, an Azure CLI or\nPowerShell command, an SDK call or a template, arrives at the same front door:\nAzure Resource Manager, which authenticates and authorises it and hands it to\nthe service that does the work. A virtual machine, a database server and a DNS\nzone are all resources in its sense, each with an ID, a type and a set of\nproperties.\n\nResources are filed in a hierarchy four levels deep. Each resource belongs to\nexactly one resource group, a folder meant for things that are deployed, updated\nand deleted together; deleting the group deletes everything in it. Resource\ngroups live inside a subscription, which is at once a billing unit and a\nboundary for access and limits. Subscriptions are gathered into management\ngroups, which can nest up to six levels and all fold up to a single root\nmanagement group for the directory. Role assignments, policies and budgets set\nat one level reach everything beneath it.\n\nOver the whole hierarchy sits identity. An organisation's users, groups and\napplications live in a Microsoft Entra ID tenant, the directory Microsoft called\nAzure Active Directory until 2023, and every Azure subscription trusts exactly\none such tenant to sign people in and decide what they may do.\n\n## Why it matters\n\nMost teams that run on Azure got there through something they already had. A\ncompany whose staff sign in to Microsoft 365 already keeps its people in an\nEntra tenant, and a subscription that trusts that tenant can give the same\naccounts access to servers and databases without a second user directory to\nmaintain. For a supplier, the choice of Azure therefore turns less on feature\ntables than on where the client's identities and data already sit.\n\nThe hierarchy is the second reason, because governance lives in it. Microsoft's\nguidance on organising resources recommends separate subscriptions for\ndevelopment, testing and production, so that environments cannot grow\ndependencies on each other and each can carry its own policies and its own line\non the bill. Azure Policy then enforces house rules at whatever level it is\nassigned. It can deny a change, such as a resource created in a region the\norganisation does not use, or log it for someone to review, and Microsoft\nadvises starting a new policy in audit mode before letting it block anything.\n\nThe third reason is that Azure can be described rather than clicked. Resource\nManager accepts declarative templates, written in JSON or in Bicep, and works\nout for itself what to create or change. Running the same template twice leaves\nthe same result, and Azure stores the state of what was deployed, so there is no\nseparate state file to look after. An estate written this way can be reviewed as\na pull request, rebuilt after a bad day and copied to a new environment by\nchanging its parameters, which is the main defence against an Azure setup that\nonly one person understands.\n\n## What it looks like in practice\n\nThe [tech stack page](/tech-stack/) names Google Cloud and AWS as the company's\nusual clouds, with Azure where a client already runs there. Two portfolio\nentries show that Azure work from both ends.\n\nOn the infrastructure side,\n[three Azure environments were written down as code](/portfolio/provisioned-azure-infrastructure-as-code/),\nall from the same Bicep templates: the api and www as Azure Container Apps on a\nmanaged environment, a PostgreSQL Flexible Server, Redis for caching, Front Door\nwith a WAF policy in front, and a VNet, NSG and private DNS underneath, with Log\nAnalytics wired in for diagnostics and images pulled from the project's Azure\nContainer Registry. Because the templates are parameterised, a new environment\nor a change to an existing one is a pull request, and drift stopped being a\nmystery, since the code is the source of truth.\n\nOn the release side,\n[deployment to Azure was owned end to end](/portfolio/owned-end-to-end-platform-deployments-to-azure/).\nEvery release travels development, staging and production in that order, and\nnobody pushes straight to a live environment. CI/CD builds and ships\nthe images, while the templates hold each Azure environment to the same\nshape, so a promoted build never lands somewhere subtly different. Settings that\nvary between environments are stored separately from secrets, which is what\nlets one built artefact pass along the whole route and pick up its own settings\non arrival, with no ad-hoc manual deploys along the way.\n\n## Where it goes wrong\n\nThe costly Azure mistakes tend to share one shape: a gap between what someone\nbelieves Azure did, or allowed, and what Resource Manager actually did.\n\n**A template deploy leaves things behind and puts things back.** Resource\nManager deploys in incremental mode by default. In that mode, a resource that\nexists in the resource group but is missing from the template is left alone, so\ndeleting a database from the template file leaves the server running and\nbilling. The same mode reapplies every property of each resource it redeploys,\nand a property the template leaves out is reset to its default. A setting\nsomeone changed in the portal is therefore overwritten on the next routine\nrelease. The what-if preview can warn of both, though Microsoft notes that its\noutput may contain noise, false positives among which a real change is easy to\nmiss. The template should state every value that differs from the default, and a\ndeployment stack, which detaches or deletes the resources that drop out of a\ntemplate, handles the leftovers.\n\n**Every environment shares one subscription.** It is the quickest way to start:\none subscription, a resource group per environment, and broad roles granted at\nthe top so nobody is blocked. Production and test then share role assignments,\npolicies, limits and a single bill, and a script aimed at the test group runs\nwith rights that also reach production. Separate subscriptions under a\nmanagement group give each environment its own access, rules and costs.\n\n**A directory admin is taken for a resource admin.** Microsoft Entra roles, such\nas Global Administrator, govern the directory: users, groups, passwords and\ndomains. Azure roles, such as Owner and Contributor, govern resources. By\ndefault neither set reaches into the other. A Global Administrator can flip a\nswitch that grants them User Access Administrator at the root scope, which lets\nthem view every resource and assign access in any subscription in the tenant.\nThe switch exists to regain access to a subscription, and Microsoft's guidance\nis to remove the elevated access once the work is done. Left in place, it is a\ndirectory account that can hand itself any resource.\n\n**A container app that never sleeps.** Azure Container Apps can scale most\napplications to zero replicas when nothing is calling them, which is part of its\nappeal for small services. Applications that scale on CPU or memory load are the\nexception and cannot scale to zero. An app given a CPU rule keeps at least one\nreplica running through every quiet night, so a service chosen for idling\ncheaply should scale on HTTP traffic or another event source.\n\n## Further reading\n\n- [Azure documentation](https://learn.microsoft.com/en-us/azure/) — Microsoft's\n  index of guides, quickstarts and references for every Azure service.\n- [What is Azure Resource Manager?](https://learn.microsoft.com/en-us/azure/azure-resource-manager/management/overview)\n  — the layer management requests pass through, and how to draw resource groups.\n- [Organize your Azure resources](https://learn.microsoft.com/en-us/azure/cloud-adoption-framework/ready/azure-setup-guide/organize-resources)\n  — the Cloud Adoption Framework on management groups, subscriptions and\n  resource groups.\n- [Deployment modes](https://learn.microsoft.com/en-us/azure/azure-resource-manager/templates/deployment-modes)\n  — what incremental mode leaves running and which properties a redeploy resets.\n- [Deployment stacks](https://learn.microsoft.com/en-us/azure/azure-resource-manager/bicep/deployment-stacks)\n  — managing a template's resources as one unit, including those removed from\n  it.\n- [Azure roles and Microsoft Entra roles](https://learn.microsoft.com/en-us/azure/role-based-access-control/rbac-and-directory-admin-roles)\n  — where directory roles end, where resource roles begin, and how elevation\n  joins them.\n\n\u003chttps://engineer.company/glossary/azure/\u003e\n","date_modified":"2026-11-12T05:34:23Z","description":"How Microsoft's rented computing is organised into tenants, subscriptions and resource groups, and where Azure deployments and permissions go astray.","formats":{"gemtext":"https://engineer.company/glossary/azure/index.gmi","gopher":"https://engineer.company/glossary/azure/gophermap","html":"https://engineer.company/glossary/azure/","json":"https://engineer.company/glossary/azure/index.json","markdown":"https://engineer.company/glossary/azure/index.md","pdf":"https://engineer.company/glossary/azure/index.pdf"},"kind":"page","language":"en","title":"What is Azure?","url":"https://engineer.company/glossary/azure/","version":1}