Hi AWS, I have created CloudWatch Dashboard for one of my HES application and it has several widgets which captures metrics like Request Count
, Request Time
, 5xx Error Count
, CPU Utilization
, Memory (RAM) Utilization
, GPU Memory Utilization
, Disk Utilization
and Status Check Failures
. The metrics are captured for every one hour time period (12 am - 1 am) and stored in the csv file.
Now the caveat is the monitoring team is updating the sheet manually and it is a pain to sit in front of machine 24*7 to evaluate the things which could have been monitored using some automation. What I was planning to do is to implement a lambda function which will capture these metrics value and store it in a CSV/Excel file and this data would be generated on scheduled basis using CloudWatch Event Rule and a notification will be sent via SNS if value for any given metric exceeds its threshold.
Am I thinking in the right direction or is there any better solution which will be faster, HA and cheap to implement with scalability in place as it will be configured for multiple applications down the line.
Please suggest some good alternatives along with the approached I am proposing.
Hi @Gary Mclean, I have added a screenshot and for example the number of
Request Count
for a given hour is 23 so I want to store this data using some form of automation in a csv file. If you can suggest whether the current pattern I am planning to use fits into that or is there any alternative which could improve this problem statement much better.