This is an automated email from the ASF dual-hosted git repository. mxm pushed a commit to branch main in repository https://gitbox.apache.org/repos/asf/flink-kubernetes-operator.git
The following commit(s) were added to refs/heads/main by this push: new 47326397 [hotfix][docs] fix some typos (#601) 47326397 is described below commit 473263978e8be1f614f4d5951c5dd76299ffa698 Author: yangjf2019 <54518670+yangjf2...@users.noreply.github.com> AuthorDate: Wed May 17 18:04:59 2023 +0800 [hotfix][docs] fix some typos (#601) --- docs/content/docs/custom-resource/autoscaler.md | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) diff --git a/docs/content/docs/custom-resource/autoscaler.md b/docs/content/docs/custom-resource/autoscaler.md index b1bdb797..748f9dc2 100644 --- a/docs/content/docs/custom-resource/autoscaler.md +++ b/docs/content/docs/custom-resource/autoscaler.md @@ -56,7 +56,7 @@ To disable scaling actions, set: `kubernetes.operator.job.autoscaler.scaling.ena Depending on your environment and job characteristics there are a few very important configurations that will affect how well the autoscaler works. -Key configuration areas +Key configuration areas: - Job and per operator max parallelism - Stabilization and metrics collection intervals - Target utilization and flexible boundaries @@ -67,9 +67,9 @@ The defaults might work reasonably well for many applications, but some tuning m ### Job and per operator max parallelism When computing the scaled parallelism, the autoscaler always considers the max parallelism settings for each job vertex to ensure that it doesn't introduce unnecessary data skew. -The computed parallelism will always be a divisor of the max_parallelism number. +The computed parallelism will always be a divisor of the max parallelism number. -To ensure flexible scaling it is therefore recommended to chose max parallelism settings that have a [lot of divisors](https://en.wikipedia.org/wiki/Highly_composite_number) instead of relying on the Flink provided defaults. +To ensure flexible scaling it is therefore recommended to choose max parallelism settings that have a [lot of divisors](https://en.wikipedia.org/wiki/Highly_composite_number) instead of relying on the Flink provided defaults. You can then use the `pipeline.max-parallelism` to configure this for your pipeline. Some good numbers for max-parallelism are: 120, 180, 240, 360, 720 etc. @@ -120,7 +120,6 @@ flinkConfiguration: kubernetes.operator.job.autoscaler.target.utilization.boundary: "0.2" kubernetes.operator.job.autoscaler.restart.time: 2m kubernetes.operator.job.autoscaler.catch-up.duration: 5m - pipeline.max-parallelism: "720" ```