Hi Mark,
I already explained that this latency is only occurring because of the
shuffle step before async IO (e.g. data is sent over network).
If you replace
val x : DataStream[String] = someIntegers.map( _ =>
s"${System.currentTimeMillis()}")
with
val x : DataStream[String] = someIntegers.shu
Hi Arvid,
The http client is not my buttoleneck as I said before I check the async
and I have a delay until it enters to asyncinvoke about 80 ms if some can
explained me why we have such big delay I have attached a sample code in my
previous email can some one explain the delay
Thanks
On Mon, 6
Hi Mark,
Async wait operators cannot be chained to sources so the messages go
through the network stack. Thus, having some latency is normal and cannot
be avoided. It can be tuned though, but I don't think that this is the
issue at hand as it should mostly impact latency and affect throughput
less
Hi Benchao,
i have run this in the code:
println(env.getConfig.getAutoWatermarkInterval)
and got 200 i do fully understand how watermarks and AsyncOperator operator
works, but
i have decided to make a simple test that should evaluate the time it takes
to enter to the asyncInvoke method and it l
Hi Mark,
could you please check if you can tune akka? Usually in async I/O, the used
library uses a thread pool that becomes the actual bottleneck.
If you configure async I/O to use a capacity of 100 and parallelism of 8 on
one node, you also need to have ~800 threads in akka (500 might be enough
Hi Mark,
According to your data, I think the config of AsyncOperator is OK.
There is one more config that might affect the throughput of AsyncOperator,
it's watermark.
Because unordered async operator still keeps the order between watermarks,
did you use
event time in your job, and if yes, what's
Hi Benchao
The capacity is 100
Parallelism is 8
Rpc req is 20ms
Thanks
On Sun, 5 Jul 2020, 6:16 Benchao Li, wrote:
> Hi Mark,
>
> Could you give more details about your Flink job?
> - the capacity of AsyncDataStream
> - the parallelism of AsyncDataStream operator
> - the time of per blocked r
Hi Mark,
Could you give more details about your Flink job?
- the capacity of AsyncDataStream
- the parallelism of AsyncDataStream operator
- the time of per blocked rpc request
Mark Zitnik 于2020年7月5日周日 上午3:48写道:
> Hi
>
> In my flink application I need to enrich data using
> AsyncDataStream.uno
Hi
In my flink application I need to enrich data using
AsyncDataStream.unorderedWait
but I am getting poor perforce at the beginning I was just working with
http call, but I have switched to grpc, I running on 8 core node and
getting total of 3200 events per second my service that I am using is no
Hi All,
I have a flow which contains few external REST calls, in few different
operators.
As I read in documentation, using AsyncFunction is the recommended way to
execute these calls.
Few questions regarding that:
1. Can someone explain better the 'capacity' parameter? How it affects
latency/
10 matches
Mail list logo