RADAR-Backend
RADAR-Backend copied to clipboard
RADAR-base
Kafka backend for processing device data
Realtime Consumers are consumers for the results from the invocation of a model from the realtime analysis pipeline (see [model-builder](https://github.com/RADAR-base/model-builder/tree/dev/model-builder), [model-invocation-endpoint](https://github.com/RADAR-base/model-builder/tree/dev/model-invocation-endpoint)and [ksqldb](https://github.com/RADAR-base/ksql-extras)). The results from the pipeline are in JSON...
Timeouts for monitor should be based on topics. Since for example, questionnaire app will be considered disconnected after a very long time compared to the passive app.
As specified here - [Kafka Stream Usage Patterns](https://cwiki.apache.org/confluence/display/KAFKA/Kafka+Stream+Usage+Patterns#KafkaStreamUsagePatterns-Howtocomputewindowedaggregationsoversuccessivelyincreasingtimedwindows?), We can maybe create successive streams for time windows since our time windows are successive in nature (10 seconds can feed into...
In source statistic monitor, the deserialization fails if schema with the id is not found. This further results in #79. We should use the `AbstractKafkaAvroDeserializer` class from radar-commons which has...
Create windowed streams for monitor topics like `android_phone_usage_event`. Also create a new windowed stream for `questionnaire_completion_log` topic. This is to make the streams compatible with the rest-api and dashboard and...
Instead of implementing source statistics as a monitor, we could implement it as a stream by using [`KStream#merge(KStream)`](https://kafka.apache.org/11/javadoc/org/apache/kafka/streams/kstream/KStream.html#merge-org.apache.kafka.streams.kstream.KStream-). This would allow using a stream StateStore to store the statistics in...
Just a question. Is this kind of behaviour normal for streams ? Here is a screen grab of htop
How to scale up partitions Current streams implementation is sensitive to partition number changes. It assumes that all messages for a given patient goes through the same partition for all...
