Why is the SLA not applying to all requests for the Time to First Response Report?

[thariea_whisker]

Please see the image for the settings for the Time to First Response SLA. When we run a report e.g. for March 2020 the number of requests being shown in the report is only two thirds of the number of overall requests for the month e.g. there were 497 service desk requests created but when we run the Time to First Response report it only shows 297. Can anyone tell me why the report is only showing some of the requests as i would expect to see the report show me the Time to First Response for all the requests as per the goals set. We have checked and all the requests are labelled correctly. 

Thanks

Thariea

[Jack Brickey · Accepted solution]

Hi @thariea_whisker there is no image. There could be any number of reasons, e.g. the way you have your filter for goals might exclude some or maybe you created the SLA after some issues were already complete. SLAs are not applied to historical issues.

[thariea_whisker]

The SLA was set prior to the beginning of March so it should bring in all requests. Hopefully the image will show this time.

[Jack Brickey]

Based on the goals I would expect the following to occur:

• all issues with a label of SFPC-04 to have an SLA that would breach at 72(h
• all issue with a label PF SFPC-03, 02 or 01 to have an SLA that breaches in 10m
• all other issue would not have the time to first response SLA

so my question is, for the issues that don’t display the SLA do they have any one off the four labels? If not then it is working properly. However if one or more has the label and no SLA then that is an issue. Please share an image of such.

[thariea_whisker]

That's correct that is exactly how it should work. We have run a report and all issues for March have been assigned with one of the labels:

As you can see below the Time to First Response only picks up 286/482 requests: The Time to First Response (below) picks up most 442/482 but not all requests:

[Jack Brickey]

It is difficult for me to troubleshoot at a distance but my theory here (current) is that the error is in the way you are crunching/assessing the results. To test this let’s simply things by moving away from these reports and into JQL queries.

create the following JQL

project = xxx and resolution is not empty and labels in (1,2,3,4) and resollutiondate >= startOfYear() and resolutiondate <= startOfYear(3M)

Hopefully I have the correct syntax here. note I shorthanded the labels here so expand. This should give you a total number of issues that have been resolved in Q1 (met and breached. How does this number compare to your Time to Resolution chart.

[thariea_whisker]

I am new to JQL, I did not get a project option so have run the query on all created service requests but it is returning 0 so I know I have done something wrong. Can you see what it is? Thanks for your help with this.

[thariea_whisker]

Also we only started using the labels mid feb so I don't expect to see numbers before then

[Jack Brickey]

Sorry. What I mean is to run this query in the Advanced search screen not in a Report.

1. click magnifying glass (assuming you are still on old navigation where it is on the sidebar)
2. click Advanced search
3. click on switch to JQL and paste the search in the search bar

one thing to point out here is that the met/breach report only show issue that have been resolved not issues still open. That may already be obvious to you.

[thariea_whisker]

I am in the new navigation view is it the search bar at the top of the new layout? thanks

[thariea_whisker]

[thariea_whisker]

[Jack Brickey]

I don’t have the new navigation but appears to be so. Now I see 804 issues listed here so something is amiss with you previously mentioned 400+ issues. Again I’m quite confident we are simply dealing with understanding what the report is showing us. Try adding “and “time to resolution” = breached()” to the JQL. This should give us the total number of breached issues in Q1.

again it is difficult to analyze this w/o seeing it firsthand.

[thariea_whisker]

re

[thariea_whisker]

I don't think the first two queries were right to show just March which is what i m looking at, the query above gave me 459 so this is inline with the Time to Resolution numbers but as we know not the Time to First Response numbers. will try adding in the time to first response met thanks

[thariea_whisker]

This is the number met which matches the Time to First Response report

Below is the number breaches which matches the Time to First Response report

So something is making some of the March requests not be included, will have to keep trying to figure it out!

[Jack Brickey]

sorry i had misread your report to be for the full Q1 but now see it is for March. please copy/paste these two JQLs and let me know the total # of issues for each (shown in the top right of your results, e.g. the 459 in your last image).

1. project = "transforming systems customer support" and labels in (SFPC-01, SFPC-02,SFPC-03,SFPC-04) AND resolutiondate > startOfMonth(-1) and resolutiondate < startOfMonth()
2. project = "transforming systems customer support" and labels in (SFPC-01, SFPC-02,SFPC-03,SFPC-04) AND resolutiondate > startOfMonth(-1) and resolutiondate < startOfMonth() AND "time to resolution" = breached()

#1 will yield all issues that were resolved during March w/ those labels

#2 will yield those issues that breached during March w/ those labels

[thariea_whisker]

459 for the first query

second query didn't return any issues

[thariea_whisker]

when i changed the second query to time to first response i get 3

[Jack Brickey]

ok so comparing to your earlier report you are showing a total of 442 issue so we are 'missing' 17. So you need to dig into your report and the queries to figure this out TBH.

I just spent 15m doing this for a project I have and I was seeing some discrepancies too. But digging in I started to figure out why, e.g. my SLAs ignore certain issue types but someone changed the issue type post-breach so I could never get things to add up.

What I suggest to you is to shrink your report time to say a week then compare the numbers to the JQL query results until you can figure out the pattern. This will allow you to go thru w/ fine-tooth-comb. Also, I note that you have broken down met/breached by priority but your first two contractual met/breached should be sufficient to focus on since they seem to sum the parts correctly.

In the end I'm confident the report is correctly displaying as it is just based upon a query. The challenge is in trying to make things match up.

[thariea_whisker]

Thanks Jack, I will have a good old audit! 

[thariea_whisker]

Sorry to post again but i have notice this, some issues are not showing the Time to First Response:

and others are as below and these ones are in the report and the above ones are not and there does not seem to be any clear reason for this, am i missing something obvious?

[Jack Brickey]

does SEH-9473 have one of the 4 labels on it AND was that label added at create time. if added after the fact it will not trigger. BTW...labels are a really bad choice for things like this as it is a free text field. I would encourage you to consider using a select field, Components, etc. and require it to be selected at creation time.

[thariea_whisker]

yes it has the label i will double check with the team if they always apply on creation, thanks for the tip think we are on to something

[Jack Brickey]

click on the history tab for SEH-9473 and see if that label added after the fact. we may have a smoking-gun here?

[thariea_whisker]

Jack, I have moved to using priority instead of labels and made this mandatory, won't know for sure if this will work until Monday but I feel confident this was the issue. You are an absolute star thanks so much for you help. Keep safe and well best wishes Thariea

[Jack Brickey]

cool. fingers crossed then we can put this novel to bed. ;-) have a great weekend!

[thariea_whisker]

:-D fingers and toes ;-)