Flush the record buffer before the timespan of all records exceeds 24 hours

[jackwakefield]

• What kind of change does this PR introduce? (Bug fix, feature, docs update, ...)

This is a bug fix for https://github.com/maxbanton/cwh/issues/96, which occurs when calling PutLogEvents with records spanning more than 24 hours.

• What is the current behavior? (You can also link to an open issue here)

The AWS SDK throws a CloudWatchLogsException with the message:

The batch of log events in a single PutLogEvents request cannot span more than 24 hours.

• What is the new behavior (if this is a feature change)?

The CloudWatch handler will check whether a record's timestamp is more than 24 hours after the earliest timestamp, and if so calls PutLogEvents before adding the new record to the buffer.

• Does this PR introduce a breaking change? (What changes might users need to make in their application due to this PR?)

No

• Other information:

N/A

[coveralls]

Coverage increased (+1.6%) to 91.346% when pulling f2f8fed90a057834e005a08a8f1b40900693b5f1 on AlsoAsked:24-hour-timespan into 6433db00611a4522f5bf4c2f3228fdaf9bd2a7a6 on maxbanton:master.