fix: issue with empty prior_values list

[danielvaughn]

Fixes ValueError when passing mode=sum to log_metric

Observed Behavior:

When you log_metric("some_metric", some_number, mode="sum"), the SDK raises the following error. This does not occur for any other aggregation modes.

Caught an error in Dreadnode. This will not prevent code from running, but you may lose data.
Traceback (most recent call last):
  File "sdk/dreadnode/task.py", line 326, in __call__
    span = await self.run(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

 [...stack trace...]
 
  File "sdk/dreadnode/metric.py", line 91, in apply_mode
    self.value += max(prior_values)
                  ^^^^^^^^^^^^^^^^^
ValueError: max() arg is an empty sequence

Resolution:

I swapped the summation method for an approach that can handle an empty list.

---

Generated Summary:

• Updated the Metric class to modify how values are aggregated in "sum" mode.
• Changed the logic from using max(prior_values) to sum([self.value, *prior_values]).
• This change enables more accurate cumulative metrics rather than only taking the maximum value.
• Potential impact includes adjustment of metric calculations which may affect downstream processes relying on these metrics.

This summary was generated with ❤️ by rigging