s one left? And if null, I instantiate a new instance
> > in my code? With billions of small events ingested per day, I can
> > imagine this to be another small performance improvement especially in
> > terms of garbage collection…
> >
> > Best regads
> >
> > T
especially in
terms of garbage collection…
Best regads
Theo
*From:*Till Rohrmann
*Sent:* Mittwoch, 19. Februar 2020 07:34
*To:* Jin Yi
*Cc:* user
*Subject:* Re: Parallelize Kafka Deserialization of a single partition?
Then my statement must be wrong. Let me double check this. Yesterday
when
…
Best regads
Theo
From: Till Rohrmann
Sent: Mittwoch, 19. Februar 2020 07:34
To: Jin Yi
Cc: user
Subject: Re: Parallelize Kafka Deserialization of a single partition?
Then my statement must be wrong. Let me double check this. Yesterday when
checking the usage of the objectReuse field, I
Then my statement must be wrong. Let me double check this. Yesterday when
checking the usage of the objectReuse field, I could only see it in the
batch operators. I'll get back to you.
Cheers,
Till
On Wed, Feb 19, 2020, 07:05 Jin Yi wrote:
> Hi Till,
> I just read your comment:
> Currently,
Hi Till,
I just read your comment:
Currently, enabling object reuse via ExecutionConfig.enableObjectReuse()
only affects the DataSet API. DataStream programs will always do defensive
copies. There is a FLIP to improve this behaviour [1].
I have an application that is written in apache beam, but
Hi Theo,
the KafkaDeserializationSchema does not allow to return asynchronous
results. Hence, Flink will always wait until
KafkaDeserializationSchema.deserialize returns the parsed value.
Consequently, the only way I can think of to offload the complex parsing
logic would be to do it in a
Hi,
As for most pipelines, our flink pipeline start with parsing source kafka
events into POJOs. We perform this step within a KafkaDeserizationSchema so
that we properly extract the event itme timestamp for the downstream
Timestamp-Assigner.
Now it turned out that parsing is currently the