Rolling window over time period

[Ockenfuss]

Hello everyone,

is there a way to do rolling window operations with a fixed-time window?
Similar to this line in pandas (from pandas rolling examples)

df_time.rolling('2s').sum()

I am aware of the existence of xarray.resample and also just discovered issue #3216.

Or do you know an easy workaround to apply rolling-window operations to unevenly sampled data, besides interpolating to an even grid?

Thank you all,
Paul