Transition to span=1m as a standard - fix memory queries overload

[Guilhem Marchand]

• Queries using span=1s have an expensive in memory cost due to the way mstats works in memory, where the full stats have to be created whenever data points exist or not

• All queries are being migrated from span=1s to span=1m to avoid long time running for large time ranges, or OMM to kill the splunkd child processes

[Guilhem Marchand]

Fix Issue #40

→ <<cset c03433bf6dd4>>