Configuration problem for Graphite

Hi, I'm having trouble configuring Graphite DB for exporting metrics. Flink version 1.20 . As you can see in the error log, the hostname is correct but the port configuration is not working and appears to be *-1*. - 2024-12-18 13:52:40,863 ERROR org.apache.flink.runtime.metrics.ReporterSetup

Re: Flink job performance

How many taskmanagers and server do you have? Can you also share the task managers page of flink dashboard? On Mon, Apr 15, 2024 at 10:58 AM Oscar Perez via user wrote: > Hi community! > > We have an interesting problem with Flink after increasing parallelism in > a certain way. Here is the sum

Delayed Window Trigger

Is it possible to trigger a window without changing window-start and window-end dates? I have a lot of jobs run in window tumble (3H) and when they are all triggered at the same time, it causes performance problems. If somehow I can delay some of them 10-15 minutes , without changing the original

GlobalWindowAggregate

Hi, Can someone tell what GlobalWindowAggregate is? it is always %100 busy in my job graph. GlobalWindowAggregate(groupBy=[deviceId, fwVersion, modelName, manufacturer, phoneNumber], window=[TUMBLE(slice_end=[$slice_end], size=[3 h])], select=[deviceId, fwVersion, modelName, manufacturer, phoneN

Bloom Filter for Rocksdb

Can someone tell the exact performance effect of enabling bloom filter? May enabling it cause some unpredictable performance problems? I read what it is and how it works and it makes sense but I also asked myself why the default value of state.backend.rocksdb.use-bloom-filter is false. We have a

Delayed Window

Hi, Is it possible to delay a window trigger without changing window-end and window-start times? Thanks

Re: Custom Prometheus metrics disappeared in 1.16.2 => 1.17.1 upgrade

Have you checked the metric changes in 1.17. >From release notes 1.17: https://nightlies.apache.org/flink/flink-docs-master/release-notes/flink-1.17/ Metric Reporters # Only support reporter factories for instantiation # FLINK-24235 # Configuring reporters by their class is no longer supported.

backpressure for just one subtask

Hi, I have a job running with parallelism=24. I see that only one subtask is %100 busy and the others are %100 idle. When I checked the received message counts, I saw that they are almost identical. How can I figure out why this task causes backpressure and why only one subtask is %100 busy. Than

Parallel Count Metrics relation

Hi, I have a job with parallelism=8. I had some performance issues and I increased the parallelism . It was ok till 10 but when I set it to 12 or something bigger I noticed that, no performance metric such as busy,backpressure or idle metrics are working on Flink UI. If I set parallelism, after w

Re: Memory Leak

; especially when your flink job uses some native library. > To diagnose such problem, you can refer to [1][2] for more details about > using NMT and jeprof. > > [1] > https://erikwramner.files.wordpress.com/2017/10/native-memory-leaks-in-java.pdf > [2] https://www.evanjone

Re: backpressured metrics doesnt work

u need to provide more context, maybe help to find the root > cause. > > Best, > Ron > > Kenan Kılıçtepe 于2023年9月4日周一 21:49写道： > >> Hi, >> >> Any idea why backpressured metrics are not working and how I can fix it? >> >> [image: image.png] >> >> Thanks >> Kenan >> >>

Memory Leak

Hi, I have Flink 1.16.2 on a single server with 64GB Ram. Although taskmanager.memory.process.size is set to 4m, I can see memory usage of the task manager exceed 59GB and OS kills it because of OOM. I check the RSS column of application top for memory usage. I don`t see any heap memory p

backpressured metrics doesnt work

Hi, Any idea why backpressured metrics are not working and how I can fix it? [image: image.png] Thanks Kenan

Re: Task Manager getting killed while executing sql queries.

Can it be a memory leak? Have you observed the memory consumption of task managers? Once, task manager crush issue happened for me and it was OOM. On Mon, Aug 28, 2023 at 9:12 PM Neha Rawat wrote: > Hi, > > > > Need some help with the below situation. If would be great if someone > could give s

Re: Uneven TM Distribution of Flink on YARN

Have you checked config param cluster.evenly-spread-out-slots ? On Mon, Aug 28, 2023 at 10:31 PM Lu Niu wrote: > Hi, Flink users > > We have recently observed that the allocation of Flink TaskManagers in our > YARN cluster is not evenly distributed. We would like to hear your thoughts > on this

Stoping a Job Without killing Task Manager

Hi, When I try to stop a job, if the job can not be stopped after a while, task manager gets killed. This is a big problem for me as there may be other jobs running on the same task manager. Is there a way to stop jobs that get stuck for some reason without causing the task manager to terminate?

Re: Request-Response flow for real-time analytics

table you can merge them. On Thu, Aug 24, 2023 at 4:10 PM Jiten Pathy wrote: > Most of the aggregates can be added/removed/updated dynamically, it would > be easier from an implementation point of view, if we could use SQL. > > On Thu, 24 Aug 2023 at 16:09, Kenan Kılıçtepe > wrot

Re: TaskManagers Crushing

d on the Kafka side at that time. > > Best, > Ron > > Kenan Kılıçtepe 于2023年8月20日周日 08:51写道： > >> Hi, >> >> I have 4 task manager working on 4 servers. >> They all crush at the same time without any useful error logs. >> Only log I can see is some disco

TaskManagers Crushing

Hi, I have 4 task manager working on 4 servers. They all crush at the same time without any useful error logs. Only log I can see is some disconnection from Kafka for both consumer and producers. Any idea or any help is appreciated. Some logs from all taskmanagers: I think first server 4 is crus

Re: Kafka Disconnection Error

Under low workload it is INFO but under heavy workload, it causes system crushes. On Wed, Aug 2, 2023 at 7:18 AM liu ron wrote: > Hi, Kenan > > I think you maybe can get help from Kafka community. IMO, it is just an > info level log, whether it has a real impact? > > >

Kafka Disconnection Error

I got a lot of these disconnection error logs. Why? My flink and kafka clusters are running in Google Cloud and I dont think there is a network issue. Also I got this error even my workload is very low. 2023-08-01 21:54:00,003 INFO org.apache.kafka.clients.NetworkClient [] - [Prod

Kafka Exception

Any help is appreciated about the exception below. Also my Kafkasource code is below. The parallelism is 16 for this task. KafkaSource sourceStationsPeriodic = KafkaSource.< String>builder() .setBootstrapServers(parameter.get( KAFKA_SOURCE_STATIONS_B

Table API and env parallelism relation

Hi, I have been experiencing a strange problem. I have been using table api and toChangelogStream() for the results. At first, even though it was running in the development environment, when I deployed my job onto the production, toChangelogStream() was not generating any results. Strangely, then