One client on your network is saturating the link — a heavy downloader, a misbehaving IoT device, or just a customer plan that's supposed to be capped at a certain speed. You need to limit bandwidth for that one address without touching anyone else's connection. MikroTik's Simple Queue is built for exactly this: a per-target speed cap you can set up in one line, no queue tree required.
Identify the target
Simple Queues are matched by target address (or subnet), so you need the client's IP first — check IP > DHCP Server > Leases if it's a dynamic address, or use a static reservation so the limit doesn't drift when the lease renews.
Create the queue
max-limit takes two values separated by a slash: upload first, download second (from the router's perspective, that's the client's upload and download).
/queue simple add name=client1 target=192.168.88.10/32 max-limit=10M/10MThat caps the client at 10 Mbps up and 10 Mbps down. Use different numbers on each side of the slash for an asymmetric plan, e.g. max-limit=5M/20M.
Mind the queue order
Simple Queues are evaluated top to bottom, and a packet is matched by the first queue whose target fits it. If you have a broad queue (say, a whole /24) above a narrower one for a single IP inside that subnet, the broad rule wins and the specific one never fires. Keep more specific targets above broader ones, the same way you'd order firewall rules.
Verify the limit is applied
Run a speed test from the client and watch the queue's counters update in real time:
/queue simple print statsYou should see the rate approach the configured cap and hold there under sustained load.
Why do it this way
Limiting per client with a target-matched Simple Queue, instead of capping the whole WAN link globally, is what actually solves the "one user is hogging the connection" problem. A global cap slows everyone down equally and does nothing to stop one client from taking more than their share below that ceiling; a per-target queue gives every client its own lane, so one heavy user can't degrade service for the rest regardless of what the total link capacity is doing. For anything past a handful of simple per-client caps — shared parent queues, prioritization between traffic types — a queue tree is the right tool, but Simple Queue covers the large majority of "limit this one client" requests on its own.
How MoniTik helps
MoniTik graphs real-time bandwidth usage per device, so instead of guessing whether a 10 Mbps cap is actually being respected under load, you can watch the traffic pattern directly and confirm the limit is doing its job — or catch the client that's still finding a way around it.