Slow Service - can it be called Downtime?

[ManP]

If an IT Service is slow, whatever be the reasons, can it be considered as downtime? Can we define like slowness of 20% or more compared to normal is considered downtime. This is what my organisation would like to define.

I disagreed. My view is unless there is a service disruption and users cannot use the IT Service it cannot be treated as downtime. Even though there is service degradation or reduced performance, but since users are continuing using the service it cannot be called downtime.

Am I right or wrong?

ManP

[Cekir]

Actually it depends on what availability is needed and is defined in SLA.
If it is the service That manages the surgery online or the flight control, then delays in service response is a very serious incident and it actually is a downtime. If however it is a CRM, the user may hold the customer or write down on the paper and then work with the application slowly. In such situation it is not to be considered as downtime. Unless the speed is critical.

In other words - If the slow down of the service makes the system not usable (remember that the business defines it) then it is a downtime.
_________________
Krzysztof (Chris) Baczkiewicz
IT Standards Support
Eracent

[UKVIKING]

Cekir is correct

If the service has an associated SLA which states that if the service has ___ periods of time whether there is service disruption of __ duration, an incident ticket is raised at ___ level so the issue can be solved

However, it aint really downtime but the way I used service degration is to equate 1 hour of severe degradation as a fraction of downtime

if the sla does not say anything about degradation in service... then the incident would be raised as a par normal for the type

It also depends on the service, the customers etc

If your customers are commercial and spending money and they cant because of the degradation...then it is a higher issue
_________________
John Hardesty
ITSM Manager's Certificate (Red Badge)

Change Management is POWER & CONTROL. /....evil laughter

[dboylan]

[Fabien]

Yes, I think the discussion is precisely on the definition of downtime.

If you limit your interpretation to whether a service is completely interrupted or not, you will end up in the sort of situation that ITIL is striving to resolve, which is a disconnect between the quality of service delivered and the customer perception of the value of that service.

A more correct definition of downtime would be the status of a service degraded to the point of failing to provide the business service it is designed for.

By this definition, you should also define that threshold in your SLAs.
_________________
BR,
Fabien Papleux

Accenture
Technology Consulting | Service Excellence
Red Badge Certified

Twitter @itilgeek

[alvinaloy]

Downtime for me, relates more to an availability issue, while slow performance would be a capacity issue.

[Jinx003]

Hi Alvinaloy,

I disagree. You might have the capacity to deliver the service. However, due to temporary malfunction, the level of service might be down or degraded. Hence, it cannot be a capacity issue.

I totally agree with Cekir's and UKVIKING's answer. Its the SLA which helps you in classifying if there is a downtime of service when you have partial service available.

Jinx003

[goitilcouk]

I concur, its all down to the SLA and what constitutes the "service".

This also expands not just to performance but service elements. Let me explain....

We have just had (and are still in) a big debate with our outsource partner re a P1 recently. The "service" is 2 x AS400's in an active passive state with real time data mirroring. As you can appreciate this is a critical system as it drives our supply chain operation.

We have a breach of the data mirroring for 36hrs which meant if we have to go to HA (ie the passive machine) our data would be 36 hours old - hence no useable service!

Our call as the customer was that their was a breach as a core component was down rendering the paid for service not available. The suppliers view was that their was no breach as the boxes were always available to the network.

It comes back to the same thing..... a clear definition of what is the service?

[GMuir]

Setting aside the SLA for the moment, the underlying question that needs to be answered is what level of slowdown would still be acceptable to the business?

I often use the phrase "functionally unavailable" in reference to this issue, it helps to link service degradation with incident concepts.

So the question becomes "At what point does the service become functionally unavailable?"