Re: [Dev] [API Analytics] - Job scheduling

2016-05-17 Thread Rukshani Weerasinha
Hi all, I will update this thread once the documentation is updated with information about the Spark scripts to be executed for this scenario. Best Regards, Rukshani. On Wed, May 18, 2016 at 9:57 AM, Nirmal Fernando wrote: > Hi Shavantha, > > Alerts are generated real-time of

Re: [Dev] [API Analytics] - Job scheduling

2016-05-17 Thread Nirmal Fernando
Hi Shavantha, Alerts are generated real-time of course and API subscribers can control whether they'd want to receive alerts or not. On Wed, May 18, 2016 at 9:49 AM, Shavantha Weerasinghe wrote: > Hi Nirmal > > Shouldn't we give the API subscriber the option to

Re: [Dev] [API Analytics] - Job scheduling

2016-05-17 Thread Sachini Jayasekara
Hi Shavantha, Alerts will be generated as and when an abnormal event comes after the learning period or after the initial script execution. After the initial script execution users will get the alerts every time an alert triggering event comes. On Wed, May 18, 2016 at 9:49 AM, Shavantha

Re: [Dev] [API Analytics] - Job scheduling

2016-05-17 Thread Shavantha Weerasinghe
Hi Nirmal Shouldn't we give the API subscriber the option to decide/control on what time lines he she wants the alerts generated. Regards, Shavantha Weerasinghe Senior Software Engineer QA WSO2, Inc. lean.enterprise.middleware. http://wso2.com http://wso2.org Tel : 94 11 214 5345 Fax :94 11

Re: [Dev] [API Analytics] - Job scheduling

2016-05-17 Thread Nirmal Fernando
Hi Iranga, Since these are data-driven alerts, IMO it is fine to send out alerts after a some time period. In fact, we have some alert types which require a learning phase, hence alerts will not get triggered during the learning phase. +1 for documenting the spark script scheduling times. On

[Dev] [API Analytics] - Job scheduling

2016-05-17 Thread Iranga Muthuthanthri
Hi, In API analytics for alert generation for use cases such as abnormal API Request count it's required to the spark scripts be executed before to get the percentile calculations. The spark scripts by default are scheduled for 12pm every day ( in this scenario) meaning at the start alerts will