Save and Start from Internal State

The Save Internal Sate and Start from Internal State functions are properties of the Feature Vector Clustering and the Reidentification microservices. Both functions are used to start the concerned microservice from a saved internal state instead of an empty database. This is useful for example in a case of unexpected stop or crash of the microservice. By default, the microservice starts with an empty database and records feature vectors and clusters while running. If the save_internal_state cluster configuration parameter is set to true (see Configuring Feature Vector Clustering) the Save Internal Sate function saves the clustering database state to a dedicated Kafka topic at frequent times. This information describes how to restore the internal state of the microservice at the given time when the last save was made. If the start_from_internal_state cluster configuration parameter is set to true (see Configuring Feature Vector Clustering) the Start from Internal State function tries to read the last saved state once from the mentioned dedicated topic. If the reading is successful the microservice continues to run with the last saved database state (as if all events happened again from the original start to the stopping of the microservice).