More expansive server metrics
Samsung is requesting the following metrics to be implemented in a future version to allow better visibility into usage and health of CircleCI Server:
- # of requested builds: the number of builds that have been requested over a specific time frame.
- # of failed/successful builds: we currently have "dead" which accounts for successful and failed builds, can this be split into failed and successful?
- # of circleci frontend access: the number of times The UI is accessed
- # of user trend: this is available through the UI and REPL, could we expose this as a metric?
- # of project trend: the number of projects building on CircleCI
- # of times GHE apis were called from CircleCI
- # of times CircleCI apis were called
- # of jobs each nomad client are executing: we have an aggregate metric returned for jobs running across the cluster. would it be feasible to split this out for each client?
Idea is in JIRA at https://circleci.atlassian.net/browse/CIRCLE-15063