Question: Unit Testing a Decorated Function with max_time about backoff

[mickey]

A bit late to the party but for future reference I both tried monkey-patching sleep and timedelta.total_seconds without success.
I finally made it work with a lookup function:

def get_mysql_max_retry_time():
   return int(os.getenv("MYSQL_MAX_RETRY_TIME", "150"))


@backoff.on_exception(backoff.expo, pymysql.err.MySQLError,
                      max_time=get_mysql_max_retry_time, jitter=backoff.full_jitter)
def ping:()
  ...

So you could probably monkeypatch this function.
In my case, I'm using dotenv so I'm not monkeypatching but declared MYSQL_MAX_RETRY_TIME=0 in .env.test.
This works great.

[bgreen-litl]

For unit testing functions with requests calls in them, I recommend responses https://github.com/getsentry/responses

With backoff in particular, if I expect the call to fail, I usually mock time.sleep() to be a no-op.

However, I think that probably doesn't work with max_time, so this is a good question. I'd be willing to add a "best practice" regarding this to the backoff documentation once we figure out what that practice should be...

[fxfitz]

So I'm already using responses for this exact test, so I'm good on that front. :-)

I actually tried mocking time.sleep() just as it's done in the backoff tests, but I land up getting an OverflowError:

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
.tox/py36/lib/python3.6/site-packages/backoff/_sync.py:109: in retry
    seconds = _next_wait(wait, jitter, elapsed, max_time_)
.tox/py36/lib/python3.6/site-packages/backoff/_common.py:38: in _next_wait
    seconds = jitter(value)
.tox/py36/lib/python3.6/site-packages/backoff/_jitter.py:28: in full_jitter
    return random.uniform(0, value)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

self = <random.Random object at 0x7fa8b689cc18>, a = 0, b = 179769313486231590...5356329624224137216

    def uniform(self, a, b):
        "Get a random number in the range [a, b) or [a, b] depending on rounding."
>       return a + (b-a) * self.random()
E       OverflowError: int too large to convert to float

.tox/py36/lib/python3.6/random.py:369: OverflowError
----------------------------------------------------------------------- Captured log call -----------------------------------------------------------------------
_common.py                  75 ERROR    Backing off get_ruleset(...) for 0.2s (RuntimeError: Unable to get latest ruleset: HTTP 404 Not Found)
_common.py                  75 ERROR    Backing off get_ruleset(...) for 1.2s (RuntimeError: Unable to get latest ruleset: HTTP 404 Not Found)
_common.py                  75 ERROR    Backing off get_ruleset(...) for 1.4s (RuntimeError: Unable to get latest ruleset: HTTP 404 Not Found)
_common.py                  75 ERROR    Backing off get_ruleset(...) for 7.3s (RuntimeError: Unable to get latest ruleset: HTTP 404 Not Found)
_common.py                  75 ERROR    Backing off get_ruleset(...) for 6.5s (RuntimeError: Unable to get latest ruleset: HTTP 404 Not Found)

[fxfitz]

FWIW, I just switched to max_tries for now. I'll leave this open though for now.