--- dead MAINPID fails fast without polling
omnigraph-server exited during startup before serving; see `journalctl -u omnigraph-server`. A fresh or never-applied storage root needs `systemctl restart omnigraph-cluster-apply` (or the bootstrap unit) before the server can start.
elapsed 0s, requests: 0
--- served equals expected with a live MAINPID is ready
elapsed 0s, requests: 1
--- served below expected fails fast
omnigraph-server serves 1 of 2 declared graphs. Graphs are opened only at startup, so waiting will not help: run `systemctl restart omnigraph-cluster-apply` (or the bootstrap unit) to apply them, then restart omnigraph-server.
elapsed 0s, requests: 1
--- a quarantined graph fails fast
omnigraph-server quarantined 1 graph(s) during startup and will not serve them until it restarts; see `journalctl -u omnigraph-server` ('graph quarantined during startup' lines) and `omnigraph-cluster status`.
elapsed 0s, requests: 1
--- no expected count accepts zero served graphs
elapsed 0s, requests: 1
--- 503 then 200 is transient
elapsed 5s, requests: 3
--- 200 that is not a readiness report fails fast
omnigraph-server answered http://omnigraph.rehearsal.invalid:8090/readyz with a body that is not a readiness report: {}
elapsed 0s, requests: 1
--- 200 with ready false keeps polling, then times out
omnigraph-server did not answer http://omnigraph.rehearsal.invalid:8090/readyz within 3 seconds
elapsed 4s, requests: 4
--- never ready within the timeout fails with the timeout message
omnigraph-server did not answer http://omnigraph.rehearsal.invalid:8090/readyz within 3 seconds
curl: (22) The requested URL returned error: 503
elapsed 4s, requests: 4