Fix progress serialization issue

[consideRatio]

Fixes jupyterhub/zero-to-jupyterhub-k8s#1501

[betatim]

In some other projects I use something like the following to serialise datetime objects (this is whats failing/why travis is unhappy):

import json

class JSONEncoder(json.JSONEncoder):
    def default(self, obj):
        if isinstance(obj, datetime.datetime):
            return obj.isoformat()
        return super().default(obj)
json._default_encoder = JSONEncoder()

but I am not sure this is "the right way" of doing this.

Reading https://code-maven.com/serialize-datetime-object-as-json-in-python which you linked in jupyterhub/zero-to-jupyterhub-k8s#1501 (comment) looks like the better option of doing the same.

What I don't know how to do is how to pass the default argument because the serialisation happens somewhere deep inside :-/

[manics]

What uses the raw-event? Would a combination of event.to_dict() and converting the datetime field to an iso8601 string (.isoformat()) work?

[consideRatio]

@clkao added it with a motivation that makes sense to me generally in #361.

[manics]

v = kubernetes.client.models.V1Event(involved_object={}, metadata={}, event_time=datetime.datetime.now())
d = v.to_dict()
for (k, v) in d.items():
    if isinstance(v, datetime.datetime):
        d[k] = v.isoformat()

[consideRatio]

manics you have the series field of the event object that van have datetimes in it as well (MicroTime ins the k8s api ref). So it must go deeper if we do this.