how to reduce update delay of network client status?

[RST]

Hi there,

to visualize the status of clients in a Meraki network a little application is cyclically (every 2 seconds) querying the API endpoint: BASE_URI/networks/NETWORK_ID/clients (https://developer.cisco.com/meraki/api-latest/#!get-network-clients)

The result is parsed and displayed in a simple GUI.

Now I observed that the displayed status has significant delay (~minutes) to the actual status of a client. In our application, some clients go offline from time to time and have to be restarted by cycling the according switch port. The affected clients are responsive on the network again, long before the status display changes back to “Online”. The responses to the API calls come every 2 seconds without noticeable delay, but the content is changing with large delay.

Since I am completely new to using the Meraki API I would be grateful for any hints to understand this behavior or to find a better approach to get “Realtime” updates of the network client status.

Thanks!

[RaphaelL · Accepted solution]

Hi ,

I think this is expected : if they are currently active, as seen below in Figure 1. The activity threshold for a client is one minute. If a client does not pass traffic for longer than one minute, then the client will no longer be considered active.

https://documentation.meraki.com/General_Administration/Cross-Platform_Content/Clients_Usage_Page_Ov...

[RST]

Thank you for the link it is very informative and explains the behaviour.

[GreenMan]

Firstly;   I wouldn't necessarily assume that, just because we have an API endpoint for a particular variable, that is what is used by the Meraki Dashboard to populate what's presented in the UI.   Just by way of example, the limit on API calls per second (currently 10) would make it impossible, I think, to use this public endpoint for this use case.   To my knowledge, there's no way to speed this up.

[RST]

Maybe I missunderstand something here but the endpoint I mentioned gives one complete list of all clients in the network. It is queried only once every two seconds. I especially used it to avoid the Rate Limit. Is this wrong?

I'm aiming for update rates of several seconds instead of minutes. I thougt this could be possible.

[GreenMan]

Sorry - re-reading your post once more I realise I totally misunderstood;   you effectively created your own mini-dashboard using the API...  I still don't know of a way to make the data you are pulling more up-to-date though;   bear in mind it's got to be pulled from the devices, collated and stored in the cloud before it can be made available.

[RST]

Thanks a lot to both of you for the fast feedback! You pointed me into the right direction.

I guess for faster status updates I'll have to use another technique like pinging the clients.