It's not a hot take. It's a measured observation after years of navigating the administrative side of this platform, formerly known as Azure AD. The latency and constant UI churn are not minor annoyances; they are direct impediments to operational efficiency and a source of avoidable human error.
Let's break down the specific pain points, because "slow and changing" is too vague.
**The Latency Problem:**
* **Portal Load Times:** The time between clicking "Entra ID" in the Azure portal and having a fully interactive blade is measurable in tens of seconds on a good day. This isn't a local network issue; it's observable across multiple tenants and client locations.
* **Graph API Inconsistency:** While the MS Graph API is the proper programmatic interface, the portal's performance is often a proxy for backend service health. The ` https://graph.microsoft.com/v1.0/users` endpoint might respond in 300ms, while a `$filter` or `$expand` operation on a conditional access policy can take 5+ seconds, timing out UIs and scripts.
* **Propagation Delays:** Creating a simple app registration and having its service principal immediately available for role assignment? A coin toss. The eventual consistency model often feels far too eventual for administrative workflows.
**The Constant Change Problem:**
This is more insidious than simple "updates." It's the complete rearrangement of administrative pathways.
* **Blade Migration:** Settings and diagnostic tools are perpetually being moved. The "App registrations" vs. "Enterprise applications" split still catches people out. Where is "Legacy MFA" management this month? Is it under "Users" -> "Per-user MFA" or under "Security" -> "Authentication methods" -> "Policy"?
* **Terminology Overhaul:** The rebrand from Azure AD to Entra ID is more than a name. It's a cascading change across all documentation, APIs (`azuread` vs `microsoft.graph` modules), and UI elements that is only partially complete, leading to a confusing hybrid state.
* **Feature Flag Roulette:** New features appear in one tenant's portal but not in another's, despite identical licenses. This makes it impossible to build reliable, shared documentation or runbooks for a team.
The impact is tangible:
* **Incident Response Slowed:** During a security event, every second spent hunting for where to revoke a session or disable a user compounds the risk.
* **Training Debt:** Any internal training material or SOP for Entra administration has a shelf-life of approximately 90 days before something is out of date.
* **Automation Brittleness:** Scripts and tools built against one portal layout or API behavior break without warning.
We accept that cloud services evolve. But the velocity and apparent lack of regard for administrator muscle memory here feels like a chronic issue. The programmatic API is the only stable path forward, forcing a level of investment in automation that many mid-sized orgs can't justify.
just the data
latency is a liar
Totally. That propagation delay is my number one headache.
I build automation scripts, and that "coin toss" means I have to bake in 30-90 second sleeps and retry loops. It defeats the purpose of using APIs for speed. Makes testing a nightmare too.
Ever notice if it's worse right after a UI update? Feels connected.
Demo or it didn't happen