wowcoders/tspark

TimeSeries Java client for Facebook Beringei. It also includes query service with tags support for metrics.

tspark

TimeSeries Java client for Facebook Beringei. It also includes query service with tags support for metrics.

Prerequisites - Download and deploy the beringei server from facebook

https://github.com/facebookincubator/beringei

Start the Beringei Server

./beringei_main -beringei_configuration_path /beringeidata/beringei.json -data_directory /beringeidata/data -port 9999 -create_directories -sleep_between_bucket_finalization_secs 60 -allowed_timestamp_behind 900 -bucket_size 1200 -buckets 144 -mintimestampdelta 0 -logtostderr -v=1

Building the Beringei Java Client

mvn clean package

This creates the following jar files

beringeiclient/target/beringeiclient-0.0.1-SNAPSHOT.jar
TSPark/target/TSPark-0.0.1-SNAPSHOT.jar
dist/target/TSPark-jar-with-dependencies.jar (fat jar)

Samples

Send & Retrieve metrics data using Java Client

Grafana dashboards

Setting up the grafana

Add data source - sample http://127.0.0.1:58080/tsdb

Installing the dashboards

Accesing the data from grafana

Importing the dashboard from the sample json file

Features

Store and Retieve Metrics to/from Beringei

Aggregations

QueryService APIs

http://<host>:<port>/tsdb/api/query
http://<host>:<port>/tsdb/api/aggregators
http://<host>:<port>/tsdb/api/suggest

(Default sample)	
http://127.0.0.1:58080/tsdb/api/query
http://127.0.0.1:58080/tsdb/api/aggregators
http://127.0.0.1:58080/tsdb/api/suggest

tspark's Issues

What is Beringei's measured put performance on a virtual machine?

My result in the ubuntu 18.04 is about 370-400 records/s. This is really not "high performance". Is there a problem with my operation? I changed minTimestampdelta to 0.