Hi,
since I don't know the implementation of the Sink I can only guess. I
would say you get 82 * 300 connections because you will get 82 instances
of a sink operator and each of those would then have a connection pool
of 300 connections. The individual sink instances will (potentially) run
on different machines and not share the connection pool.
Best,
Aljoscha
On 05.10.20 22:28, Vijay Balakrishnan wrote:
HI,
Basic question on parallelism of operators and ConnectionPool to DB:
Will this result in 82 * 300 connections to InfluxDB or just 300
connections to InfluxDB ?
main() {
sink = ....createInfluxMonitoringSink(..);
keyStream.addSink(sink).addParallelism(82);//will this result in 82 * 300
connections to InfluxDB or just 300 connections to InfluxDB ?
}
private ..... createInfluxMonitoringSink(...) {
final OkHttpClient.Builder okHttpClientBuilder = new
OkHttpClient.Builder()
.readTimeout(timeout, TimeUnit.MILLISECONDS)
.connectTimeout(timeout, TimeUnit.MILLISECONDS)
.writeTimeout(timeout, TimeUnit.MILLISECONDS)
.connectionPool(new ConnectionPool(300, 60,
TimeUnit.SECONDS));
try (InfluxDB influxDB = InfluxDBFactory.connect
(host, userName, pwd, okHttpClientBuilder)) { ......}
}
TIA,
