[
https://issues.apache.org/jira/browse/SDAP-537?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Kevin Marlis updated SDAP-537:
------------------------------
Description:
Dynamically scaling ingester pods on an as needed basis (within the confines of
available node resources) is possible, but not with Kubernetes HPA.
[KEDA|[https://keda.sh/]] looks to be an open source autoscaler that works with
external metrics, like RMQ queue size.
We would need to explore the maturity of KEDA (they claim to be production
grade) and see if it's a good fit for our needs or if there's better
alternatives.
was:
Dynamically scaling ingester pods on an as needed basis (within the confines of
available node resources) is possible, but not with Kubernetes HPA.
[KEDA|[https://keda.sh/]] looks to be an open source autoscaler that works with
external metrics, like RMQ queue size.
We would need to explore the maturity of KEDA (they claim to be production
grade) and see if it's a good fit for our needs or if there's better
alternatives.
> Dynamic ingester scaling based on RMQ queue size
> ------------------------------------------------
>
> Key: SDAP-537
> URL: https://issues.apache.org/jira/browse/SDAP-537
> Project: Apache Science Data Analytics Platform
> Issue Type: New Feature
> Reporter: Kevin Marlis
> Priority: Minor
>
> Dynamically scaling ingester pods on an as needed basis (within the confines
> of available node resources) is possible, but not with Kubernetes HPA.
> [KEDA|[https://keda.sh/]] looks to be an open source autoscaler that works
> with external metrics, like RMQ queue size.
> We would need to explore the maturity of KEDA (they claim to be production
> grade) and see if it's a good fit for our needs or if there's better
> alternatives.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
