Minion Agent replaces SQL Server Agent with something built for a whole fleet, not one box -- orchestration, alerting, and access control that scale to hundreds of servers from a single console. Here's the short list of what sets it apart, with a link to the full story on each.

Run jobs across your entire fleet, from one place

Native SQL Server Agent only ever runs a job on the box it lives on. Minion Agent orchestrates jobs across every server you register, from a single console -- see Target Servers for how a server gets added to the fleet in the first place.

Import your existing SQL Agent jobs in minutes

Point Minion Agent at a live server (or paste the script SSMS already generates) and it parses every job, step, and schedule, flags anything it can't confidently translate, and lets you fix those before committing -- nothing is created until you say so. See Importing from SQL Server Agent.

Gates: pause or stop a run on a real condition, not just a schedule

A file showing up, a resource threshold, another job's outcome -- a Gate checks it and decides whether a run starts, pauses, or stops, live, not just on a timer. See Gates.

Jobs that survive an Availability Group failover automatically

Point a job or step at "whichever server currently holds the Primary role" instead of a fixed server name, and it keeps working through a failover with nobody touching anything. See Availability Group Settings.

Chain jobs by real outcome, not by guessing at timing

A Trigger starts one job the moment another actually finishes with the status you picked -- Succeeded, Failed, whatever you need -- through the same queue a manual Run uses. See Job Triggers.

Alerts that respect a schedule and look the way you want

Route alerts differently for Business Hours versus After Hours without touching the rule itself, and build the exact subject/body you want with reusable Message Styles instead of one fixed template. See Alert Schedules and Message Styles.

An audit trail that's enforced, not just possible

Every mutating call, from every app that talks to Minion Agent -- including your own scripts -- is required to record who changed what, or the request fails outright. There's no way to add a new feature that quietly skips the audit log. See Activity (Audit Log).

Permissions down to a single job or step

Grant a role access fleet-wide, or scope it to one target server, one folder, one job, even one step -- not an all-or-nothing switch. See Roles & Permissions.

See exactly who can touch what, before it's a problem

Pull a full report of every permission a specific user holds, or every user who can touch a specific role, instead of piecing it together by hand. See Permissions Reports -- By User and By Role.

Block dangerous commands before they ever run

A fleet-wide banned-phrase list (plain text or regex) is checked against every step's command text and blocks a match by default -- with per-object overrides for the rare case that actually needs one. See Command Blacklist.

Know which credentials have more access than they should

Scan your target servers for credentials sitting in elevated roles like sysadmin, so "which login can do way more than it needs to" is an answer you can pull up, not a guess. See List Credentials and Request an Elevated Role Scan.

One value, one place, used everywhere

Define a Constant once and reference it from any step's command text -- change it in one place instead of hunting through every job that uses it. See Constants.

Pull diagnostics from your whole fleet in one click

Collect logs and diagnostics from every server in the fleet in a single request instead of connecting to each one by hand -- a peer that can't be reached is noted, not a blocker for the rest. See Fleet Support Bundle.

A full REST API -- script anything the Console can do

Every screen in the Console calls the exact same API you can call yourself, in C# or PowerShell, with the exact same permissions model. See Calling the API From Your Own Code and the full API Guide.

Fleet-wide password policy and account control

Set what a generated app-account password has to look like once, fleet-wide, and disable an account instantly across every sign-in method -- Windows, Azure AD, or app account -- with an already-open session cut off immediately. See Password Policy and Account Lifecycle.

See All Help Topics for the complete list, the Quick Start Guide to get going, or Release Notes for what's changed lately.