The possibility of using interaction-based Genesys controls...

[tony]

Appropo nothing at all - I'm just an advocate of "out of the box" thinking... 

I've been looking closely at what can be achieved using Alarm Detection/Condition/Reaction within SCS/SCI and it came to me that you [i]could [/i] control certain aspects of your Genesys environment remotely - by phone (or email - or any other interaction method)...

1. There are settings you can use within [i]Alarm Detection [/i] that can detect entries in any Genesys Application Log File, including the potential to recognize a specific attached data item (KVP) in the TServer Log... [i]<can you see where this is going...? >[/i]
2. There are [i]Alarm Conditions [/i] you can set on a particular Application instance....
3. There are [i]Alarm Reactions [/i] which can invoke all sorts of things - including Application Setting chnages, OS commands and other types of on-line and ff-line scripts (SNMP)...

So, it should be feasible to;

4. Create an IVR Script or Routing Strategy loaded on an external DN which, given the right User input(s) would create a specific User Data (KVP) entry in a TServer Log File.
5. Create an Alarm Detection based on the TServer Log File User Data (KVP) item.
6. Create an Alarm Reaction to Load and Unload various items, including Routing Strategies and changes to Sub-Routines.
7. Call the external DN an enter digits (via a Dialogic or IVR) to apply the KVP within the TServer Log File which, with [i]Alarm Detection [/i] could enable a specific [i]Alarm Reaction [/i] to stop/start/change setitngs within the Genesys environment.

Like I said - appropo nothing but the concept is quite interesting, if you need to make a remote change (frequently) on the fly and you don't have immediate access to your environment... Perhaps if you are "on-call" and it is your responsibility to invoke operational/reactive changes to your Genesys Applications or Routing on the fly, from home...

Any thoughts on this...?  Or have I finally gone mad....? 

Tony

[ecki]

Sounds awesome, though I haven't heard about this yet. What version are you talking about? 7.6 or 8?

[tony]

I don't think it's particularly version specific.  I've just never taken a really good look at the potential available within the Management Layer... When I did, it occured to me that many, many wonderful things may be possible and, being the sort of gent I am, I thought I would share those thoughts...

[SisB]

We have been using a similar approach. One example: If there is any config issue, an alarm is raised from strategy, which in turn creates a ticket or sends out an e-mail through reaction script.

After taking care of the config, operations call up a number, enter alarm id/specific cancel event flow to clear the alarm.

But, how do you detect a KVP from detection script? I thought you can detect only some system variables/SNMP parameters through it.

Currently we detect INTERESTING strings/KVP in logs through custom/shell scripts.

Thanks.

[tony]

Lemme take a look - I found it earlier today... I'll take a look again, tomorrow...

T

[tony]

OK - I can see that this would be difficult - but not impossible...

I think the only issue is within the [i]Alarm Detection[/i], which is constrained to Host and Application CPU/MEM errors [b]but [/b] it can also detect SNMP Traps...

To me, it is perfectly feasible to create a Strategy which monitors all Interactions coming through an RP, for a KVP condition to be met (e.g. DNIS="+44123456789" or through the application of a KVP for CollecDigits on an IVR...)  Once the condition is met, an SNMP Trap can be invoked from within IRD (unfortunately, SNMP is licensed seperately by Genesys! ) and it is that which is recognized by a pre-defined [i]Alarm Detection[/i], set at regular intervals within SCI  From there, it would be a simple case of setting up an SCI [i]Alarm Condition [/i] and an associated [i]Alarm Reaction [/i], which can then invoke pre-defined changes within the Genesys Applications.

I [i]think [/i] this is what you are saying you already do, SisB...?

Tony

[SisB]

Tony,

we have something similar but rather than monitoring through strategy (we do throw alarms from strategy whenever needed and easily possible  ), we grep the logs and send traps. It gives bit more flexibility as we can monitor several other parameters (or, anything we like to check in the logs).
E.g.
ASAI messages/seconds, crv utilization (Avaya)
PegDef key, no of calls in progress in T memory
etc etc.

Thank you.

[Dionysis]

Why not just use URS to trigger a web service using config SDK to achieve the same thing?

SCI seems like the "long way round" to me.

[tony]

SNMP costs, SDK costs... I suppose it's down to cost..? Mine is/was a concept but, as with most things, someone has already been and done it!

Tony

[Adam_W]

I was just thinking along the same lines as Tony while trying to solve something and I stumbled across this thread!

It would be so useful to be able to use KVPs to trigger an alarm and reaction in SCS.  I've noticed the Alarm function in IRD before but think that this can only be used for SNMP traps.

Is there really no other simple way to trigger an alarm in SCI from a strategy, without using SNMP or the Config SDK?  Seems crazy!