Hello,
We currently have multiple sites with a mix of equipment. Every site has a pair of MX250s connecting to some kind of stacked core. Our aim is to migrate everything to meraki but many sites still have other brands of switches atm.
We also have a bit of a mix of DHCP scope locations. Some on MX's, some on switch Cores and some on Domain Controllers.
I am trying to tidy up and move all the scopes to a standard location. We were thinking to put everything on the MX's, however I was wondering how much they can handle in terms of system performance?
If we do this, a single MX device is going to be doing all the firewall/content filter decisions, all the routing, traffic shaping, all the DHCP requests, site to stie VPN, maybe client VPN in the future.
How do I tell how much is too much? Some of our bigger sites can have hundreds of users and thousands of devices.
Is there a way to check its performance? see how hard its working?
Or cant they just deal with whatever we throw at it and don't worry about it?
Thanks!
Solved! Go to solution.
>Also, how do I tell what our MX's are actually doing?
The only way to get this info is to open a support ticket. However, you are not likely to hit this limit. You would need to have around 10,000 users to get near that.
There is not enough information to answer your question directly, so instead let me point you in the direction of the sizing guide. It has performance guidelines.
https://documentation.meraki.com/MX/MX_Sizing_Information/MX_Sizing_Principles
cheers! its very helpful, but im still not quite sure how I asses what load our devices are under.
First of all, the doc states maximum recommended values for each feature - but is that if that's the only feature in use? Can a device cope if its doing EVERYTHING near maximum values?
Also, how do I tell what our MX's are actually doing?
For example -
It is important to understand the number of flows, or open sessions, supported by each appliance. For purposes of sizing, a flow is any transmission on an open socket within the last 5 minutes. Note that this is not a recommended flow capacity number, but instead these values are to be maximums.
MX67 | MX68 | MX75 | MX85 | MX95 | MX105 | MX250 | MX450 | |
Maximum Concurrent Sessions | 25,000 | 25,000 | 50,000 | 125,000 | 200,000 | 250,000 | 500,000 | 1,000,000 |
So it seems we can handle 500,000 "flows" in 5 min. Great, but are we operating close to that? over it?, no where near?
The nearest info I can find is in "traffic analytics" which can only give me a 2 hour timeframe and divides the flows up per application.
I could export to CSV to combine the flows for a total over 2 hours and get SOME idea but its not exactly precise is it?
Are there other performance/activity tools im missing somewhere? Would the MX alert if it was getting over-worked?
cheers.
>Also, how do I tell what our MX's are actually doing?
The only way to get this info is to open a support ticket. However, you are not likely to hit this limit. You would need to have around 10,000 users to get near that.
sweet, thanks very much!
>Also, how do I tell what our MX's are actually doing?
I thought of a better answer. Go to Organization/Summary Report. Down the bottom left is a table showing the utilisation of the top MXs.