Open scotthellman opened 7 years ago
We currently store the Validation in the corresponding Result. Because we use multiprocessing when running Alarmageddon, all objects in the Result must be pickleable.
As of right now, no Publishers take strong advantage of the fact that the whole Validation is available when publishing. PagerDutyPublisher does the most with it, in that it creates a string of the Validation's __dict__.
To provide users with more freedom in how they create their own Validations, we should not store the whole validation in the result, and instead just store the information needed by our Publishers.
We currently store the Validation in the corresponding Result. Because we use multiprocessing when running Alarmageddon, all objects in the Result must be pickleable.
As of right now, no Publishers take strong advantage of the fact that the whole Validation is available when publishing. PagerDutyPublisher does the most with it, in that it creates a string of the Validation's __dict__.
To provide users with more freedom in how they create their own Validations, we should not store the whole validation in the result, and instead just store the information needed by our Publishers.