rate calculation

example-rate.py

def delta_sn(a, b):
    d1 = datetime.strptime(a, '%Y-%m-%dT%H:%M:%SZ')
    d2 = datetime.strptime(b, '%Y-%m-%dT%H:%M:%SZ')
    diff = (d2 - d1).total_seconds() / 60
    return round(diff, 1)


def calculate_rate(attribute_name):
    a0 = ds8[(ds8[attribute_name] > 0)]
    a0_max = ds8[(ds8[attribute_name] == ds8[attribute_name].max())]

    a01 = a0.reset_index(drop=True)
    a02 = a0_max.reset_index(drop=True)
    print('{:<12s}{:>12s}{:>20s}{:>1s}'.format(attribute_name,'duration (h)',':', str( round(delta_sn(a01['date'][0], a02['date'][0]) / 60,1)) ))
    print('{:<12s}{:>12s}{:>13s}{:>1s}'.format(attribute_name,'rate clusters/hours',':', str( round(a02[attribute_name][0] / (delta_sn(a01['date'][0], a02['date'][0]) / 60),1)) ))

    #    print(attribute_name + ' duratiom (h) :' + )
    #    print(attribute_name + ' rate clusters/hours :' + str( round(a02[attribute_name][0] / (delta_sn(a01['date'][0], a02['date'][0]) / 60),1) ))

run_date='6/21/2021'
title='1K Provision, WAN Emulation On, RHACM 2.3/AI, Bare Metal/virtual Bare Metal' + ', ' + run_date
l={
    "value" : "# clusters",
    "date" : ""
    }

fig = px.line(ds8,
              x='date',
              y=['initialized',
                 'iso_gen',
                 'iso_bmh',
                 'booted',
                 'discovered',
                 'provisioning',
                 'install_failed',
                 'completed',
                 'managed',
#                  'agents_available'
                ],
              labels=l)


fig.add_trace(go.Scatter(x=ds8['date'], y=ds8['initialized'] - ds8['completed'] - ds8['install_failed'], name='concurrency', 
                         line=dict(color='black', width=2, dash='dash')))

# calculate average concurrency, manually bound the data set 
# calculating the concurrecy without bound does not give meaningful results
cc = ds8[(ds8['date'] > '2021-06-19T02:00:00Z') & (ds8['date'] < '2021-06-19T04:00:00Z')]

cc_mean = np.mean(cc['initialized'] - cc['completed'] - cc['install_failed'])
cc_max = np.max(cc['initialized'] - cc['completed'] - cc['install_failed'])

print('approx concurrency mean                    :' + str( round(cc_mean,1) ))
print('approx concurrency max                     :' + str( round(cc_max,1) ))




calculate_rate('initialized')
calculate_rate('completed')
calculate_rate('managed')

fig.update_layout(
    title=title,
    legend_orientation='v'
)


fig.show()