The SocorroCollector has the ability to vet crashes as the come into the system. Statistically, it is not necessary to save and examine every single crash that comes into the system. Although, the system can save 100% of the crashes, in a system with a high volume, it is unlikely that there would be enough storage space.

Crashes come into the system in part called a JSON/dump pair. The JSON file contains meta information about the crash, while the dump is actually the binary dump created by Breakpad.

Throttling, the disposition of a JSON/dump pair, is controlled by the contents of the JSON file. The JSON files are collections of keys and values. SocorroCollector can examine these key/value pairs and assign a pass through probability. For example we may want to pass 100% of all alpha or beta releases to processing. In production, however, we may want to only save 10%.

For details on how to configure throtttling, see the configuration section of SocorroCollector.

enhancement proposal

Right now, the system is only able to make throttling decisions based on single values within the JSON file. Since our instance of Socorro is handling multiple products, making throttling decisions based on 'version' alone doesn't make much sense. There needs to be a reworking of the SocorroCollector throttling system to allow for switching based on any arbitrary combination of key/value pairs.