inconsistency between resample and plotting

What is your issue?

I believe there is an inconsistency between using the resample function (left labeled) and the plotting function (center labeled).
Maybe the example below help illustrating the issue:

d = [
    [1,0,0,0,0,0,1,0,0,0,0,0],
    [0,0,1,1,0,0,0,1,1,1,1,0],
    [0,0,0,0,0,0,0,0,0,0,1,0],
]

t = pd.date_range('2000-01-01 00:05','2000-01-01 00:17',freq='1t',closed='left')

da = xr.DataArray(d,dims=['i','t'],coords={'t':t})

f,axs = plt.subplots(4,sharex=True,constrained_layout=True)

def _plot(da_,ax_):
    return da_.plot(ax=ax_,add_colorbar=False,vmin=0,vmax=1.001,levels=11)

for ax,tt in zip(axs[1:],['2t','3t','4t']):
    da_ = da.resample({'t':tt}).mean()
    gr = _plot(da_,ax)
    ax.set_title(f'resample: {tt}')

_plot(da,axs[0])
axs[0].set_title('orginal')
    

for ax in axs:
    ax.grid()
    ax.set_xticks(t)
    ax.set_xticklabels(t.strftime('%M'))
    ax.set_xlim(pd.to_datetime('2000-01-01 00:00'),pd.to_datetime('2000-01-01 00:20'))
    
f.colorbar(gr,ax=axs) 

In the example above, the most relevant problem is the high value at min 11 in panel 1 that after resampling to 4 minutes in panel 4 gets shifted and displayed between minutes 6 and 10.
I know that i can shift the results from resample with the parameter loffset (='30s' ). but this that not help since now the high value in panel 1 (min 11) is also shifted to minutes 6-10 in panel 4

d = [
    [1,0,0,0,0,0,1,0,0,0,0,0],
    [0,0,1,1,0,0,0,1,1,1,1,0],
    [0,0,0,0,0,0,0,0,0,0,1,0],
]


t = pd.date_range('2000-01-01 00:05','2000-01-01 00:17',freq='1t',closed='left')

da = xr.DataArray(d,dims=['i','t'],coords={'t':t})

f,axs = plt.subplots(4,sharex=True,constrained_layout=True)

def _plot(da_,ax_):
    return da_.plot(ax=ax_,add_colorbar=False,vmin=0,vmax=1.001,levels=11)

for ax,tt in zip(axs[1:],['2t','3t','4t']):
    da_ = da.resample({'t':tt},loffset='30s').mean()
    gr = _plot(da_,ax)
    ax.set_title(f'resample: {tt}')

_plot(da,axs[0])
axs[0].set_title('orginal')
    

for ax in axs:
    ax.grid()
    ax.set_xticks(t)
    ax.set_xticklabels(t.strftime('%M'))
    ax.set_xlim(pd.to_datetime('2000-01-01 00:00'),pd.to_datetime('2000-01-01 00:20'))
    
f.colorbar(gr,ax=axs) 

Start by running the issue's reproducible example with xarray DataArray.resample and plot, then compare the resampled coordinates with the plotted positions for the 2t, 3t, and 4t cases. The work is done when the resampling and plotting behavior is made consistent or the expected coordinate behavior is clearly documented, with coverage for the demonstrated case.