java - Where to aggregate raw data analytics within microservice architecture -
we need aggregate analytics our backoffice. analytics specs crossed on our system each microservice create it's own raw data. question should aggregate raw data analytics purposes: (i thought 2 ways)
we should create new service (e.g analytics-service) aggregate raw data services(by push data it) , prepare aggregations after having backoffice taking final results
each service aggregate it's own analytics , our backoffice send requests (and cache) aggregated results microservices ?
thanks.
there ton of frameworks , services out there you. dropwizard, statsd, new relic, app dynamics. research. i'd don't roll own because you'll spend more time creating , adding features metrics system gaining value it. use out of box , running host of features developed.
-- update --
here things i'd when deciding on cross service metrics generation / collection system.
- minimal no source code manipulation. lots of libraries can inspect runtimes without modifying source code. there apis can use in conjunction instrumentation add specific metric gathering code if need later on.
- easy collection. pointing metrics gathering services collector should easy configure.
- easy viewing / aggregation. once have metrics collected, should easy see data came from, , data points are.
basically, choose 1 gets minimal work. may find base solution meet needs. if not, spend time adding in collectors / dashboards.
Comments
Post a Comment