We need to do Android -> Cloudwatch metrics so we can monitor and page (ie. no more than a 5 minute delay in data except for android devices that are offline). In servers in java, micrometer.io is a common solution with a GCP or AWS plugin to report the 4 things we are interested in PER ENDPOINT...
- request per minute
- failure per minute (we can graph % success rate from #1 and #2)
- p99 latency (the 99th slowest request out of 100 requests every minute showing us 99% of users experience this latency or better)
- failure by exceptionName per minute showing us the most important exceptions to fix first.
This is pretty standard per endpoint. We have a mesh of android devices and need to know their view of the world (the p99 view of 'most' devices that is). We would like in cloud watch (or GCP monitoring, etc), to see those 4 graphs and to be able to page on them.
I don't think 'anything' exists for this next requirement out there but just in case... In the future(maybe 1 year), we also want to be able to go offline for 5 hours and then slowly fill the last 3 hours of metrics (ie. 3 hour running window while offline). This gives us a view into offline mode performance as well.
We would love to just start with basic metrics on android. thanks, Dean