The actual requirement is if CPU/memory increases(trending up) continuously due to a memory leak and finally one day container may blast and recreate right due to memory/CPU exhausted Before blasting the container user should monitor the status by the alerts
1.increasing continuously without decreasing for period of time --> Alert should raise
2.increasing and decreasing for period of time --> Alert should not raise
Both points should include in single expression in the Prometheus rule for both CPU/Memory in single expression
My goal is to monitor container memory/cpu for a continuous increase over time and alert only if there are no decreases within a specified period