Priority based escalation

[cmadoery]

As far as i understand there are two ways to escalate.
1. the queue based escalation
2. the SLA based escalation
I actually do the SLA based way at the moment. But now we need to have a PRIORITY based escalation.
Whats best way to do that?
- Changing queue on behalf of priority?
- fiddle around with generic agent?
- ???
any suggestions are very welcome.

Christof

[crythias]

Priority is priority, how urgent to start, SLA is SLA, when to finish.

A service has attached a [business] criticality and (maybe) an SLA. That service's SLA is always (supposed to be) the same SLA, when you're working on that service. The priority is intended to be independent of that.

The criticality - impact - priority matrix assigns flag colors based upon how many people this business-critical service's downtime impacts, but it still is the same SLA: No matter what happens to the email server (service), the SLA for that is still 1 hour for first response, 2 hours for followup, etc.

I can understand the possibility of changing the SLA attached to the service based upon the priority, but this is counter to how ITIL is designed. Nonetheless, you might be able to use Generic Agent Event to change SLA based upon priority change.

[cmadoery]

totally agree.
I perfectly fulfill the SLA with the criticality - impact - priority matrix. Cool!

Still need to inform Account Mgmt. after some "update time" is over the limit. Which works perfect with queues. Also cool!
But they had the idea that OTRS shout act different on low, normal, high priority.

Still thinking about where to solve that best.

Many thanks!

[crythias]

OTRS shou[ld] act different on low, normal, high priority.

I really understand this (often requested) point of view, but the gist is still what can the field actually *do* and essentially ... nothing really but colors.

The response I have is, generally, that a customer might (constantly) initiate a request at priority 1... all that customer's requests are urgent/highest priority. So that sets an SLA, right? Except this particular request is about a 404 on a website. I can offer the user a variation of SLAs to choose from for a given service instead of priority. That way I get an idea of what he/she'd agree to the service level on it. It's also a (different) visual indicator of expectation levels. Yes, I understand you (customer) consider this a high priority, but considering what service you're asking about, I really can [continue] focus[ing] on more business-critical services while observing the response level you require/agree upon to address your issue.

Also, it allows me to reduce the priority to what I need it to be without changing the agreed response level.