Evaluation infrastructure for testing real-time content search
Systems and methods provide an experimentation system, or testing engine, for a real-time search infrastructure. An example method includes generating a snapshot of a production search environment and performing testing of a signal or index change in the snapshot. The change can be specified as parameters passed to the system. The method may include estimating an impact of the change based on the testing and determining, based on the estimate, whether the impact is positive. Responsive to determining the impact is positive the method may include generating a holdback environment of the production search environment, the holdback environment being a portion of the production search environment selected not to receive the change, pushing the change to the production search environment, and monitoring the change by running partial production traffic through the holdback environment.
1. A method comprising:
generating a snapshot of a production search environment, the search environment using a plurality of signals to generate a search result provided to a client;
performing testing of a signal or index change in the snapshot;
estimating an impact of the change based on the testing;
determining, based on the estimate, whether the impact is positive; and
responsive to determining the impact is not positive:
generating crowd-source tasks based on testing,
calculating a metric based on results of the crowd-source tasks,
determining, based on the metric, results are positive, and
responsive to determining the results are positive:
generating a holdback environment of the production search environment, the holdback environment being a portion of the production search environment selected not to receive the change,
pushing the change to the production search environment, and
monitoring the change by running partial production traffic through the holdback environment.
2. The method of claim 1 , wherein the snapshot includes a partial capture of production data.
3. The method of claim 1 , wherein the snapshot includes a partial capture of production data that represents a pre-defined number of days.
4. The method of claim 1 , wherein the testing is AB testing.
5. The method of claim 4 , further comprising, responsive to determining the impact is not positive:
determining, based on the metric, results are not positive;
providing the change to an experimentation system; and
responsive to determining results from the experimentation system are positive:
generating the holdback environment,
pushing the change to the production search environment, and
monitoring the change by running partial production traffic through the holdback environment.
6. The method of claim 1 , wherein a change affecting more than 2% of traffic is not positive.
7. The method of claim 1 , wherein the change is a new index field for retrieval.
8. The method of claim 7 , wherein generating the snapshot includes:
receiving a map from a document to the new index field;
extracting the information from the document according to the map; and
generating an index in the snapshot that includes the new index field.
9. The method of claim 1 , wherein the testing includes a query set of a pre-defined number of queries that are provided from the production search environment.
10. A system comprising:
at least one processor,
memory storing instructions that, when executed by the at least one processor cause the system to perform operations including:
generating a snapshot of a production search environment, the search environment using a plurality of signals to generate a search result provided to a client;
performing testing of a signal or index change in the snapshot;
estimating an impact of the change based on the testing;
determining, based on the estimate, whether the impact is positive; and
responsive to determining the impact is not positive:
generating crowd-source tasks based on testing,
calculating a metric based on results of the crowd-source tasks,
determining, based on the metric, results are positive, and
responsive to determining the results are positive:
generating a holdback environment of the production search environment, the holdback environment being a portion of the production search environment selected not to receive the change,
pushing the change to the production search environment, and
monitoring the change by running partial production traffic through the holdback environment.
11. The system of claim 10 , wherein the snapshot includes a partial capture of production data.
12. The system of claim 10 , wherein the snapshot includes a partial capture of production data that represents a pre-defined number of days.
13. The system of claim 10 , wherein the testing is AB testing.
14. The system of claim 13 , the memory further storing instructions that, when executed by the at least one processor, further cause the system to perform operations including, responsive to determining the impact is not positive:
determining, based on the metric, results are not positive;
providing the change to an experimentation system; and
responsive to determining results from the experimentation system are positive:
generating the holdback environment,
pushing the change to the production search environment, and
monitoring the change by running partial production traffic through the holdback environment.
15. The system of claim 10 , wherein a change affecting more than 2% of traffic is not positive.
16. The system of claim 10 , wherein the change is a new index field for retrieval.
17. The system of claim 16 , wherein generating the snapshot includes:
receiving a map from a document to the new index field;
extracting the information from the document according to the map; and
generating an index in the snapshot that includes the new index field.
18. The system of claim 10 , wherein the testing includes a query set of a pre-defined number of queries that are provided from the production search environment.