[ZBX-16601] Problem time and SLA is calculated differently for Services at specific display periods Created: 2019 Sep 05  Updated: 2024 Apr 10  Resolved: 2019 Dec 18

Status: Closed
Project: ZABBIX BUGS AND ISSUES
Component/s: Frontend (F)
Affects Version/s: 4.0.13rc1, 4.2.7rc1, 4.4.0alpha3
Fix Version/s: 4.0.16rc1, 4.4.4rc1, 5.0.0alpha1, 5.0 (plan)

Type: Problem report Priority: Trivial
Reporter: Sergejs Olonkins Assignee: Sergejs Olonkins
Resolution: Won't fix Votes: 0
Labels: SLA
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Attachments: GIF File incorrect_problem_time.gif    
Issue Links:
Sub-task
Team: Team D
Sprint: Sprint 58 (Nov 2019), Sprint 59 (Dec 2019)
Story Points: 1

 Description   

Problem description: Problem time and SLA is calculated differently for Services based on the same trigger at specific display periods if the service uptime period differs for these services. For example:
Service #1 (based on trigger 1) has uptime period set to "Monday 14:00 - Monday 16:00" and service #2 (based on trigger #1) has uptime period set to "Monday 14:00 - Monday 15:00".
Currently the service is in status problem. For period in ['Today', 'This week', 'This month', 'This year', 'Last 24 hours'] the SLA and problem time values of the two services are almost equal, but if period is set to "Last 7 days", "Last 30 days" or "Last 365 days" - problem time of service #1 has problem time 2 times less than service #2.

Steps to reproduce:
1. Create a host, a zabbix trapper item and a trigger for this item, with severity higher that "Information".
2. Navigate to Configuration => Services and create 2 services with SLA calculation based on the previously created trigger:

  • For service #1 set the uptime period for 1 hour (current time must be part of this period).
  • For service #2 set the uptime period for 2 hours (current time must be part of this period).

3. Set the trigger in status problem.
4. Navigate to Monitoring => Services and make sure that the "Problem time" has started to increase for both services.
5. Change the period to be displayed - set all of the values one by one.

Result: For some of the periods the problem time and the SLA values are almost equal for the two services (today, this week, this month and last 24 hours), but for some periods for service #1 they are two times larger than for service #2 (this year, last 7 days, last 30 days, last 365 days).

Expected: Such difference of behavior should not be observed - we either take into account the length of the uptime period, or we don't.

Example:



 Comments   
Comment by Sergejs Olonkins [ 2019 Dec 18 ]

After a deeper analisys it was concluded that the behaviour described in this task is expected.
Uptime of both services was configured to take into account problems on Mondays during a certain period of time. In case of 'Today', 'This week', 'Last 24h' and 'This month' (because the issue was reported on the 1st Monday of the month) the data that is gathered during previous Mondays is outside of the selected period and, therefore, has no impact on the calculated SLA. This explains why during these periods SLA of both services is equal.
But periods 'Last 7 days', 'Last 30 days', 'Last 365 days' and 'This year'actually do include the data gathered during previous Mondays, when there was no issue. Because the length of the uptime periods of service #1 is two times longer than of service #2, the data gathered during previous Mondays had direct impact on the obtained SLA value.

Generated at Fri Mar 14 13:31:12 EET 2025 using Jira 9.12.4#9120004-sha1:625303b708afdb767e17cb2838290c41888e9ff0.