We previously had a case openerd for the maintenance windows not working.
We noticed yesterday that some notifications were still being sent despite a maintenance window being active.
To handle out of maintenance schedules that are coming from maintenance service pod, a work around is added in NAS to have maintenance schedules cache for the deleted maintenance schedules along with with active schedules. If a deleted schedule exists , then even though active schedule exists, NAS will not add the alarm into maintenance. Currently from maintenace schedule when the schedule is extended, instead of getting EDIT schedule, DELETE and ADD schedule payload is sent to NAS. Due to having delete schedule cache. NAS is not adding the alarm into maintenance.
Fix: Handling EDIT payload at both maintenance service pod and NAS pod will add more regression and require more changes. Also, maintenance service pod is now sending the maintenance schedules to specific partition ( Based on tenantId and scheduleId ) quite for some. So, delete cache is not required now and also will solve the EDIT schedule problem (which is easyto fix now to handle the current scenario). So, it is decided to remove the delete maintence schedule cache and the corresponding checks with delete cache while processing the alarm in NAS.