Hi, This is a separate question.
As I know, the API Manager is integrating Carbon Metrics [1] and the timers are used to measure the time duration for handlers. A timer in Metrics is a duration histogram and it has predefined set of percentile values (95th, 99th etc). Will we be able to make use of metrics for alerting as well? For this, we may need to study how the reservoir works in Metrics Histogram. [1] https://github.com/wso2/carbon-metrics On Mon, Feb 15, 2016 at 4:39 PM, Sachini Jayasekara <sachi...@wso2.com> wrote: > Hi all, > > I am working on following two alert generation scenarios, > > 1. Abnormal response time/ backend time - If an API response time or the > backend time lies outside the Xth percentile. > > 2. Abnormal request counts - Count the number of requests during last > minute and check if it lies outside the Xth percentile. > > Initially, we thought of having a summarization job to calculate the > percentile values daily. However, in DAS, spark doesn't have OOB support to > calculate percentile even though hive has that as a built-in aggregate > function. > > With DAS 3.0.1 we can define UDFs (User Defined Functions), but the > function 'percentile' is a user-defined aggregation function (UDAF). > Hence, we cannot use that approach as well. > > Is there an alternative way in DAS to define custom aggregation function? > > > Thanks, > Sachini > > -- > > > > *Thanks & Regards,Sachini JayasekaraSoftware Engineer; **WSO2 Inc. * > > *lean . enterprise . middleware | http://wso2.com <http://wso2.com> * > > Mobile : +94712371165 > > _______________________________________________ > Dev mailing list > Dev@wso2.org > http://wso2.org/cgi-bin/mailman/listinfo/dev > > -- Isuru Perera Associate Technical Lead | WSO2, Inc. | http://wso2.com/ Lean . Enterprise . Middleware about.me/chrishantha Contact: +IsuruPereraWSO2 <https://www.google.com/+IsuruPereraWSO2/about>
_______________________________________________ Dev mailing list Dev@wso2.org http://wso2.org/cgi-bin/mailman/listinfo/dev
