This makes sense when you think about the inelastic demand for gas and the physical problem of storing it. The availability of storage makes gas prices change slowly most of the time. But, if there's a chance of the tanks running dry, prices go sharply up. And if there's a chance of the tanks running out of space, prices go sharply down.

In those cases they shouldn't go up immediately only during refilling an order.

Not the individual gas stations; I mean the storage capacity of the whole system.

That price drop was more the result of purchase demand drying up overnight with everyone staying home (a car not driven needs no fillups). When the local retailer goes from moving 1,000 gal per day to moving 100 gal per day (numbers made up) that incentivizes the retailer to drop their street prices to try to generate extra demand at the point of sale.

The other price increases/reductions being discussed are when the price of crude takes a dramatic swing, one sees the street prices swing asymmetrically. Crude goes up by $40/barrel today, for one day, street prices for gasoline goes up tomorrow and remains up. Crude drops by $50/barrel next week and stays down, it takes weeks before the street price of gas falls to reflect the reduced price of crude. For these scenarios, the local sales demand for gasoline at the local station would have remained relatively flat. The retailer might have seen sales swing from 1,000 gal/day to 950 gal/day after the increase (again, made up numbers), but not enough swing to impact his pricing choices.