[jira] [Created] (FLINK-7539) Make AvroOutputFormat default codec configurable

2017-08-27 Thread Sebastian Klemke (JIRA)
Sebastian Klemke created FLINK-7539: --- Summary: Make AvroOutputFormat default codec configurable Key: FLINK-7539 URL: https://issues.apache.org/jira/browse/FLINK-7539 Project: Flink Issue Ty

Re: Serialization Issue

2017-08-27 Thread Alexx
My stream is producing records of type *Tuple2* *.toString()* output *(usr12345,{"_key":"usr12345","_temperature":46.6})* where the key is *usr12345* and value is *{"_key":"usr12345","_temperature":46.6}* The *.print()* on the stream outputs the value correctly: *(usr12345,{"_key":"usr12345","_temp

Serialization Issue

2017-08-27 Thread Alexandru Gutan
My stream is producing records of type Tuple2 *.toString()* output *(usr12345,{"_key":"usr12345","_temperature":46.6})* where the key is *usr12345* and value is *{"_key":"usr12345","_temperature":46.6}* The *.print()* on the stream outputs the value correctly: *(usr12345,{"_key":"usr12345","_tem

Streaming job gets slower and slower

2017-08-27 Thread Aparup Banerjee (apbanerj)
We have a flink job with 7 subtasks. See graph below. This is on flink 1.2. [cid:image001.png@01D31EC7.FB4B5620] Here each source task consumes from a kafka topic. Data rate is low around 70-80 messages per sec. What we are noticing is after running for 2 hours or so the source tasks starts sho