Does Jira upgrade from version 9 to 11 impacts SLA calculation? My reports SLA goals calculations shows 500% and 1000%. Any suggestions.
Did you already did a Reindex or SLA recalculation (for specific items) after the upgrade? Also check the application logs for SLA or indexing errors since the upgrade.
Hi Manas,
Mark's reindex and recalculate idea is a good start. I'd look closely at the percentages themselves before you run anything, though. Atlassian's KB on how the SLA goals report is calculated (https://support.atlassian.com/jira/kb/how-sla-goals-report-vs-sla-success-rate-report-is-calculated-in-jira-service-management/) gives the goal percentage as SUCCESSFUL / TOTAL, where the successful issues are part of the total. By my reading of that formula it can't go over 100%, so a 500% or 1000% result points to broken SLA data or an index problem rather than a real figure.
The Missing SLAs KB (https://support.atlassian.com/jira/kb/missing-slas-in-jira-service-management/) walks through the recalculation, and I'd go in this order:
1. On a few affected issues, open the SLAs section, choose More actions (…) > Recalculate all SLAs, and then check whether your report numbers look normal again.2. If that fixes them and you have many affected issues, the bulk route is the destructive reconstruct call at /rest/servicedesk/1/servicedesk/sla/admin/task/destructive/reconstruct/jql, which needs the dark feature sd.sla.fix.timeline.enabled first. Take a database backup first, keep your JQL to one project, and try it on staging before production. I'd avoid the force variants entirely. The bulk steps are in https://support.atlassian.com/jira/kb/how-to-bulk-recalculate-missing-slas/.
If your numbers still look wrong after that, the exact 11.x version you're on would help narrow it down.
Hi! Atlassian's documentation for the built-in SLA goals report gives successful ÷ total, where successful issues are a subset of the total. So that report shouldn't be able to go over 100%. Values like 500% or 1000% usually point to a custom report series or to bad SLA data after the upgrade. Some things to check:
If the values still look wrong, open a ticket with Atlassian Support (Data Center includes it) and attach the debug output.
Refs:
https://support.atlassian.com/jira/kb/how-to-troubleshoot-slas-in-jira-service-management/
https://support.atlassian.com/jira/kb/how-sla-goals-report-vs-sla-success-rate-report-is-calculated-in-jira-service-management/
Well, the only place it is wrong is in in-built "SLA goals report". Rest it is correct in in-built "Custom report > SLA Success rate" and in one of my dashboard created using "Advanced Issue Filter Formula Gadget". Additionally, the SLA cycle runs fine in tickets
SLA goals report
Hey Manas, thanks for checking all of that, because it changes my read quite a bit. If your tickets and your success rate report are both right, the SLA data itself looks fine, so I'd drop the reconstruct I suggested and you don't need to take that risk. The one documented difference is that the goals report also counts SLAs that are still running, while the success rate report only counts finished ones. Atlassian's KB puts it as "on the SLA goals the total will include all of the issues where the SLA is still running/ongoing whether it is breached or not" (https://support.atlassian.com/jira/kb/how-sla-goals-report-vs-sla-success-rate-report-is-calculated-in-jira-service-management/), and that still can't push it past 100%, so it looks like something specific to the goals report. I'd go with Mark and Mertcan's full reindex next, since the stack trace in https://jira.atlassian.com/browse/JSDSERVER-16560 shows that page querying the index, though that's my reading and it isn't documented. If you still see over 100% after that, I'd raise it with Atlassian Support with your exact version and the screenshot, as I couldn't find a published bug for it. Which 11.x version are you on? I'm around if you want a second pair of eyes on it
Jira version: v11.3.12 Checking what the DBs were doing during the background reindexing. On prod, it was between 19:19 and 05:29, slight increase in CPU load on the DB (16 CPUs). On test, it was between 19:19 and 01:46, increase in CPU load on the DB (2 CPUs) - a flatish 40% extra.
It looks like you're new here. Sign in or register to get started.