This may seem like a stupid question, but I can't seem to find the answer. It would make sense to me for to only include currently online clients, but I want to be sure.
We have a branch with an MX68, which has a client recommendation of up to 50 clients. At the moment of writing this post, this location has 122 clients devices in the network-wide > clients section. However when I filter to online clients it has only 32. This should be fine right? As long as the online clients never reach more than 50, the MX68 should be able to handle the traffic fine, right?
I ask because we are getting complaints that everything is really slow at this location. We are trying to figure out what is causing it. This is our biggest branch, so I had the thought that an MX68 might not be enough for it. With there being 122 including offline, I suppose it's possible that it goes over 50 clients at some point that I'm not monitoring it. But I don't want to buy a MX84 (200 client recommendation) and it not solve anything.
Solved! Go to solution.
That number 50 is not a hard stop. It's a guideline.
Aaron Willette (SE at Meraki) has actually written a great blog post for MX sizing:
https://www.willette.works/meraki-mx-sizing/
Looking at your numbers I honnestly think you're fine.
50 active clients.
how is the device performance, that shows up at summary report?
https://meraki.cisco.com/lib/pdf/meraki_whitepaper_mx_sizing_guide.pdf
Enough WAN bandwidth? does the mx show high latency or packet loss ?
Thanks, makes sense, but I wanted to be sure. To answer your questions:
Device utilization over the past 30 days is almost never over 25%, there's one time in that history where it got close to 50%, but looks like it was just a short peak. So that means the MX should be fine, if I understand correctly.
Here is what it says for Uplinks:
Label | Loss Rate% | Latency (ms) | RTT min (ms) | RTT max (ms) | Jitter (ms) | % Time Up |
Internet 1 | 0 | 22.872 | 11.51 | 3201.798 | 1.949 | 100 |
Internet 2 | 0 | 7.983 | 7.812 | 834.297 | 0.289 | 99.99 |
Interent 2 is the primary and active WAN. Internet 1 is just a backup slow internet meant to just keep the location online if the other internet goes down. So the latency there makes sense to me.
That number 50 is not a hard stop. It's a guideline.
Aaron Willette (SE at Meraki) has actually written a great blog post for MX sizing:
https://www.willette.works/meraki-mx-sizing/
Looking at your numbers I honnestly think you're fine.
Okay cool, thanks for the help. We can rule that out for sure then.
As others said, it's active people. Meraki keeps a record of clients it's seen on your network for a while. I think for clients it's 31 days, but I'm having a hard time finding that in an actual doc beyond the API documentation.
We generally look at the status report performance, as you did. So we've got some folks at my MSP with high active client counts, but good performance scores on MX60s. We leave them alone.
Other people, I'm trying to figure out how their MX60-64s haven't melted. 🙂
Haha, that's actually pretty cool. You ended up saving money on a smaller MX. I have no idea how you could plan for that without a meraki already there though.
Network survey can help identify active client numbers. My particular company drops Auvik in to find as many things as possible, but there's a lot of available tools! We just use Auvik to monitor MSP client networks so... two birds, one stone. Two polar bears, one stone?