Influx showstopper, this is as far as I got

Alltough Influx allows irregular time series, downsampling does not work
really: if you query a time span it may omit a measurement from
result, if now data point was encountered in this span.
Likewise, if no measurement occured in the earliest grouped by
intervali(s) empty fields will be returned.
There is frustrated discussion in
influxdata/influxdb#6878 and no progress since ~6 years.

Will evaluate QuestDB in a side branch.

aquaPi/machineroom/misc_nodes.py

@@ -311,6 +311,16 @@ def query(self, node_names, start=0, step=0):
         log.debug('  done, overall %fs', time() - qry_begin)
         return result
 
+def query_influx(db_name, flux):
+    breakpoint()
+    if not INFLUX_SRV:
+        pass
+    elif INFLUX_SRV == 'localhost':
+        os.system(f'influx -database={db_name} -precision=s -execute "{flux}"')
+    else:
+        _curl_query(flux, option="db={db_name}&precision=s")
+    return []
+
 
 # ========== miscellaneous ==========
 
@@ -362,9 +372,10 @@ def get_history(self, start, step):
         return self.db.query(self._inputs.sender, start, step)
 
     def get_history(self, start, step):
-        h = influx_query('aquaPi',
-                         'select time, %s from %s where time > %d'
-                         % (','.join(self._inputs.sender), self.id, start))
+        h = query_influx('aquaPi',
+                         f'SELECT time, * FROM {self.id} WHERE time > {start}'
+                         #'SELECT time, mean(*) FROM {self.id} WHERE time > {start}'
+                         #% (','.join(self._inputs.sender), self.id, start))
         hist = []
         return hist