Schema inference and log data validation system
Systems and methods are described for generating metrics from log data items, automatically inferring one or more schemas based at least in part on analyzing samples of the log data items, validating samples of the log data items against the one or more schemas to detect log data item errors, and analyzing the log data item errors according to metrics analytics rules to determine an effect of the log data item errors on a quality measurement of the metrics.
1. A computer-implemented method comprising:
generating metrics from log data items;
automatically inferring one or more schemas based at least in part on analyzing samples of the log data items;
validating the samples of the log data items against the one or more schemas to detect log data item errors; and
analyzing the log data item errors according to metrics analytics rules to determine an effect of the log data item errors on a quality measurement of the metrics.
2. The computer-implemented method of claim 1 , wherein the log data items represent user interface (UI) interactions by a user of an application program running on a production computing system and the metrics measure performance of the UI interactions by the production computing system.
3. The computer-implemented method of claim 2 , comprising automatically inferring the one or more schemas when a new release of the application program is deployed to the production computing system.
4. The computer-implemented method of claim 1 , wherein the generating the metrics, automatically inferring the one or more schemas, validating the samples of the log data items, and analyzing the log data item errors are performed on a non-production computing system.
5. The computer-implemented method of claim 1 , wherein the log data items comprise semi-structured JavaScript object notation (JSON) objects containing collections of key-value pairs.
6. The computer-implemented method of claim 1 , comprising validating the samples of the log data items against the one or more schemas to detect the log data item errors on a periodic basis.
7. The computer-implemented method of claim 1 , comprising generating a violation map summarizing each type of log data item error and a number of the log data items having the type of log data item error.
8. The computer-implemented method of claim 1 , comprising sending an alert when the effect of the log data item errors on the quality measurement of the metrics causes the quality measurement to fail to meet a predetermined threshold.
9. An apparatus comprising:
a performance metrics generator to generate metrics from log data items;
a schema inferencer to automatically infer one or more schemas based at least in part on analyzing samples of the log data items; and
a log data validator to validating the samples of the log data items against the one or more schemas to detect log data item errors.
10. The apparatus of claim 9 , comprising validating samples of the log data items against the one or more schemas to detect log data item errors on a periodic basis.
11. The apparatus of claim 9 , wherein the log data items comprise semi-structured JavaScript object notation (JSON) objects containing collections of key-value pairs.
12. The apparatus of claim 9 , wherein the log data items represent user interface (UI) interactions by a user of an application program running on a production computing system and the metrics measure performance of the UI interactions by the production computing system.
13. The apparatus of claim 12 , comprising the schema inferencer to automatically infer the one or more schemas when a new release of the application program is deployed to the production computing system.
14. The apparatus of claim 9 , comprising:
a metrics analyzer to analyze the log data item errors according to metrics analytics rules to determine an effect of the log data item errors on a quality measurement of the metrics.
15. The apparatus of claim 14 , comprising the metrics analyzer to generate a violation map summarizing each type of log data item error and a number of the log data items having the type of log data item error.
16. The apparatus of claim 14 , comprising the metrics analyzer to send an alert when the effect of the log data item errors on the quality measurement of the metrics causes the quality measurement to fail to meet a predetermined threshold.
17. A non-transitory machine-readable storage medium that provides instructions that, if executed by one or more processors, are configurable to cause the one or more processors to perform operations comprising:
generating metrics from log data items;
automatically inferring one or more schemas based at least in part on analyzing samples of the log data items;
validating the samples of the log data items against the one or more schemas to detect log data item errors; and
analyzing the log data item errors according to metrics analytics rules to determine an effect of the log data item errors on a quality measurement of the metrics.
18. The non-transitory machine-readable storage medium of claim 17 , wherein the log data items represent user interface (UI) interactions by a user of an application program running on a production computing system and the metrics measure performance of the UI interactions by the production computing system.
19. The non-transitory machine-readable storage medium of claim 18 , comprising instructions to automatically infer the one or more schemas when a new release of the application program is deployed to the production computing system.
20. The non-transitory machine-readable storage medium of claim 17 , wherein instructions to generate the metrics, automatically infer the one or more schemas, and validate the samples of the log data items are performed on a non-production computing system.