Arvados

There is a great deal of tooling for monitoring/graphing/alerting on Prometheus, whereas our health check API is entirely Arvados specific. It would be useful to be able to expose the heath checks as Prometheus metrics so they can be consumed by the monitoring infrastructure.