I noticed in the opentsdb.conf file for 2.4.0RC2, which is bundled as part of MEP 4.0.0, that there are several entries at the foot of the file related to consuming data from streams:
# Config to enable reading from streams
tsd.default.usestreams = false
#tsd.default.consumergroup = metrics
#tsd.streams.path = /var/mapr/mapr.monitoring/streams
#tsd.streams.consumer.memory = 4194304
#tsd.streams.count = 2
#tsd.streams.autocommit.interval = 60000
#tsd.mode = ro
I guess this represents new functionality, coded specifically by MapR folks?
Is it working in the MEP 4.0.0 rpm, and if so:
- What do the settings mean? Some are self explanatory, but others are not.
- What format messages should a producer generate: "put ..." (i.e. telnet API), or JSON (i.e. HTTP API)?
- In the case of the latter (JSON), what happens to the response content?
Also, does the MEP build of OpenTSDB include the capability to direlcly load blobbed data, reported as useful for data migration in this article: Loading a Time Series Database at 100 Million Points Per Second | MapR ?