cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
1175
Views
0
Helpful
7
Replies

Monitor mode type app in Finesse?

I'd like to create a server-side app that will monitor all agents and log them out after 4 hours in a single state.

 

Is there sufficient Finesse API for this?  Based on docs I've seen, I'd have to either create a supervisor for each team or shadow logins for each agent.

 

Thanks,

Chris

1 Accepted Solution

Accepted Solutions

dekwan
Cisco Employee
Cisco Employee

Hi,

 

As you are already aware (based on your statements), Finesse is not a monitoring application and doesn't have the ability to monitor all agents.

 

If you want your app to be a listener, then what you have said of creating a supervisor for each team or shadow logins for each agent is a workaround. If you do the shadow logins, do know that it counts as an additional agent against your license as well as increase the activity on the Finesse system. With this method, you would get a notification every time the agent changes its state. So your application can probably store the timestamp when you get the notification and check periodically which agents have exceeded the 4 hr mark.

 

If this is the only scenario you are trying to accomplish, what if your server side application just polled all of the agent's state every x minutes? Does the agent inactivity have to be exactly 4 hours? The caveat for this route is that you would need to use the administrator's credentials. What your application would do is, as an administrator, it would loop through the agent's state using the GET User API and check the <stateChangeTime> tag to see when the agent changed into that state. If it is >= 4 hours, then log them out. If not, wait for the next poll. GET's are pretty lightweight, but of course it shouldn't happen every 5 seconds.

 

Thanx,

Denise

View solution in original post

7 Replies 7

dekwan
Cisco Employee
Cisco Employee

Hi,

 

As you are already aware (based on your statements), Finesse is not a monitoring application and doesn't have the ability to monitor all agents.

 

If you want your app to be a listener, then what you have said of creating a supervisor for each team or shadow logins for each agent is a workaround. If you do the shadow logins, do know that it counts as an additional agent against your license as well as increase the activity on the Finesse system. With this method, you would get a notification every time the agent changes its state. So your application can probably store the timestamp when you get the notification and check periodically which agents have exceeded the 4 hr mark.

 

If this is the only scenario you are trying to accomplish, what if your server side application just polled all of the agent's state every x minutes? Does the agent inactivity have to be exactly 4 hours? The caveat for this route is that you would need to use the administrator's credentials. What your application would do is, as an administrator, it would loop through the agent's state using the GET User API and check the <stateChangeTime> tag to see when the agent changed into that state. If it is >= 4 hours, then log them out. If not, wait for the next poll. GET's are pretty lightweight, but of course it shouldn't happen every 5 seconds.

 

Thanx,

Denise

I love this idea - thanks Denise! I can use api/Users and then loop through with api/User/XXXXX - once each hour is close enough since I can always use stateChangeTime to schedule specific queries.
Now, with multiple PGs across the enterprise, would I need to query HDS to determine all the API endpoint hosts (Finesse servers)?

Hi,

 

Each Finesse pair only knows about the agents in that particular PG(s) that it is connected to. So if you have multiple Finesse pairs in your enterprise, then you would need to do it for each one of these pairs.

 

Thanx,

Denise

Hi Denise,
Yes, that was the basis of my 2nd question. It looks like I can use the UCCE API's Machine Inventory endpoints to get all the Finesse PGs (hopefully these are addresses or host names).  It's possible this is dependent on the architecture and might not contain ALL the Finesse pairs I need. In other words, is there a top-level place I can go to discover all the Finesses without creating a custom list that'd have to be maintained?

Chris

Hi Chris,

 

My expertise is more on the Finesse side, but the Machine Inventory API is the only one I know about and I'm pretty sure that it'll give you the address/hostname. If it only gives you the address, you can always get the hostname from the Finesse SystemInfo API if you really needed it.

 

CCE doesn't have too many REST APIs out there and most of it is for configuration.

 

Thanx,

Denise

Denise,

 

Can you feed back to the BU - that it is critical to have more than one user account on Finesse that has admin privileges.

A custom Finesse Desktop will need to use the admin account - as would many bespoke apps such as below you are describing.

 

But anyone from the operational team can accidentally lock out the admin account when logging into the GUI and hence "break" the custom app from working. 

To unlock the finesse admin account - you have to change its password and then reset it back to the original one.

 

Hence the requirement for at least 2 seperate admin accounts.

One for apps and one for GUI.

 

Also we have found that we often have to give the business access to Finesse Admin - so a number of GUI admin accounts would be ideal!

 

Also it would be good to see if and when Finesse will have an all event listener feature.

It is a necessity to get this on the roadmap.

It could be caveated that when used max capacity of finesse is reduced by X % 

 

Regards,

Gerry

Hi Gerry,

 

Sure, I will forward your feedback to the Finesse PO.

 

Yes, I agree that it is good to have more than one administrator. 

 

As far as an all event listener feature, the PO is already aware of this request. At this time, Finesse is intended to be an agent desktop versus a monitoring application. Although not ideal and/or easy, at this time, you can get the events via the CTI Protocol. I am not sure where this feature request falls in the backlog, but like I said, I will forward your requests.

 

Thanx,

Denise