Create getAllEvents probe

[russmiles]

The probe's signature will likely look like:

getAllEvents: string

And simply return the full json in a string to be used to enrich an experiment's journal. This probe is unlikely to be useful to steady-state hypothesis work.

The underlying API call will implement:

> curl --request GET \
  --url '.../api/events?from=1554143400000&to=1554237055206' \
  --header 'authorization: apiToken ********' 

[russmiles]

The work will include updating the usage in the README, and also providing a short sample experiment with its usage incorporated in a samples directory.

[russmiles]

Parameters would be a window of time (start, end) and an optional delay parameter, defaulted to 5 seconds, as Instana's model can be up to 5 seconds behind the feeds from the target system.

[russmiles]

I'd like to refine this method so that it is called get_all_events_in_window as typically a window of time needs to be passed to retrieve a subset of all the events available. The signature would look like the following:

get_all_events_in_window(from, to)

I'd also like to propose that the from and to parameters could be defaulted to None, allowing for a default window of "last 5 minutes from time probe was called" to be employed.

Finally it would also be possible to support only from being specified, to defaulting to None, so that a window "from X ago to time probe was called" could be employed.

As tyo on its own doesn't seem too useful, it is suggested the above be achieved without default values and instead with function overloading?

[Lawouach]

Python doesn't really have the concept of overloading functions. You do use default values indeed.

You could achieve this with singledispatch but I'm afraid this will make the code harder to read in this case.

I would stick with keeping a single function with default values.

Note also, from is a keyword in Python, you won't be able to call your variable that way.

Finally, might be worth using dateparser to allow users to say "2 minutes ago" which reads more fluently.

start_dt = dateparser.parse(start, settings={'TO_TIMEZONE': 'UTC'})

start is a string such as "2 minutes ago" or "now" that is translated into a datetime instance in UTC timezone. If you need to serialize it back to an ISO string:

isostr = start_dt.isoformat()