fafo/jeol-t330a
15
0
Fork 0
Code Issues 4 Pull requests Projects Releases Packages Wiki Activity Actions

More succbone metrics #14

Merged
q3k merged 1 commit from metrics into main 2024-11-09 22:20:38 +00:00
Conversation 6 Commits 1 Files changed 1 +46 -2
1 changed files with 46 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files
Showing only changes of commit 0aba323779 - Show all commits

48
succbone/succd/http.go
View file

@ -163,16 +163,60 @@ func (s *webServer) viewStream(w http.ResponseWriter, r *http.Request) {
}
}
func boolToFloat(b bool) float32 {
if b {
return 1.0
} else {
return 0.0
}
}
// httpMetrics serves minimalistic Prometheus-compatible metrics.
func (s *webServer) viewMetrics(w http.ResponseWriter, r *http.Request) {
// TODO(q3k): also serve Go stuff using the actual Prometheus metrics client
// library.
// TODO(q3k): serve the rest of the data model
state := s.d.snapshot()
mbar := state.piraniMbar100.mbar
// sem_pressure_mbar is meant to represent the fused pressure value
q3k commented 2024-10-06 14:46:12 +00:00
Outdated
Copy link

Let's wait until we have more data sources and work on the model then - I don't think there's much sense in doing this ahead of time?

Let's wait until we have more data sources and work on the model then - I don't think there's much sense in doing this ahead of time?
rahix commented 2024-10-07 05:47:40 +00:00
Outdated
Copy link

My main motivation here was to not break the existing metric but keep consistency in the naming - a sem_pressure_volts would not make sense but having sem_pressure_mbar alongside sem_pirani_volts is also not nice. So by duplicating the metric, the names are consistent and we don't break already recorded metrics data for no good reason.

My main motivation here was to not break the existing metric but keep consistency in the naming - a `sem_pressure_volts` would not make sense but having `sem_pressure_mbar` alongside `sem_pirani_volts` is also not nice. So by duplicating the metric, the names are consistent and we don't break already recorded metrics data for no good reason.
// from all data sources once we have more vacuum sensors in the
// system. sem_pirani_mbar is just the reading from the pirani gauge.
fmt.Fprintf(w, "# HELP sem_pressure_mbar Pressure in the SEM chamber, in millibar\n")
fmt.Fprintf(w, "# TYPE sem_pressure_mbar gauge\n")
fmt.Fprintf(w, "sem_pressure_mbar %f\n", mbar)
fmt.Fprintf(w, "sem_pressure_mbar %f\n", state.piraniMbar100.mbar)
fmt.Fprintf(w, "# HELP sem_pirani_mbar Pressure reading by the Pirani gauge, in millibar\n")
fmt.Fprintf(w, "# TYPE sem_pirani_mbar gauge\n")
fmt.Fprintf(w, "sem_pirani_mbar %f\n", state.piraniMbar100.mbar)
fmt.Fprintf(w, "# HELP sem_pirani_volts Voltage output from the Pirani gauge, in volts\n")
fmt.Fprintf(w, "# TYPE sem_pirani_volts gauge\n")
fmt.Fprintf(w, "sem_pirani_volts %f\n", state.piraniVolts100.avg)
fmt.Fprintf(w, "# HELP sem_pirani_failsafe_active Whether pirani gauge failsafe mode is active (boolean)\n")
fmt.Fprintf(w, "# TYPE sem_pirani_failsafe_active gauge\n")
fmt.Fprintf(w, "sem_pirani_failsafe_active %f\n", boolToFloat(state.safety.failsafe))
rahix marked this conversation as resolved Outdated
q3k commented 2024-10-06 14:49:32 +00:00
Outdated
Copy link

Probably should be suffixed with _enabled or _active (same for DP lockout).

Probably should be suffixed with `_enabled` or `_active` (same for DP lockout).
fmt.Fprintf(w, "# HELP sem_dp_lockout_active Whether diffusion pump lockout is active (boolean)\n")
fmt.Fprintf(w, "# TYPE sem_dp_lockout_active gauge\n")
fmt.Fprintf(w, "sem_dp_lockout_active %f\n", boolToFloat(state.safety.highPressure))
rahix marked this conversation as resolved Outdated
q3k commented 2024-10-06 14:51:11 +00:00
Outdated
Copy link

If we rename this to sem_pump_diffusion_... and sem_pump_roughing_... we could then have aggregate metrics on eg. regex sem_pump_.*_running.

If we rename this to `sem_pump_diffusion_...` and `sem_pump_roughing_...` we could then have aggregate metrics on eg. regex `sem_pump_.*_running`.
fmt.Fprintf(w, "# HELP sem_pump_diffusion_running Whether the diffusion pump is running (boolean)\n")
fmt.Fprintf(w, "# TYPE sem_pump_diffusion_running gauge\n")
fmt.Fprintf(w, "sem_pump_diffusion_running %f\n", boolToFloat(state.dpOn))
fmt.Fprintf(w, "# HELP sem_pump_roughing_running Whether the roughing pump is running (boolean)\n")
fmt.Fprintf(w, "# TYPE sem_pump_roughing_running gauge\n")
fmt.Fprintf(w, "sem_pump_roughing_running %f\n", boolToFloat(state.rpOn))
rough, high := state.vacuumStatus()
fmt.Fprintf(w, "# HELP sem_vacuum_rough_reached Whether a rough vacuum has been reached (boolean)\n")
rahix marked this conversation as resolved Outdated
q3k commented 2024-10-06 14:52:04 +00:00
Outdated
Copy link

passed in comment, reached in metric name - let's settle on one term?

`passed` in comment, `reached` in metric name - let's settle on one term?
fmt.Fprintf(w, "# TYPE sem_vacuum_rough_reached gauge\n")
fmt.Fprintf(w, "sem_vacuum_rough_reached %f\n", boolToFloat(rough))
fmt.Fprintf(w, "# HELP sem_vacuum_high_reached Whether a high vacuum has been reached (boolean)\n")
fmt.Fprintf(w, "# TYPE sem_vacuum_high_reached gauge\n")
fmt.Fprintf(w, "sem_vacuum_high_reached %f\n", boolToFloat(high))
}
func (s *webServer) viewRoughingPumpEnable(w http.ResponseWriter, r *http.Request) {