[
https://issues.apache.org/jira/browse/MESOS-7170?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16023906#comment-16023906
]
Eric Chung commented on MESOS-7170:
-----------------------------------
We're having the same issue: our production cluster typically runs 30k+ tasks,
making the time to download the entire list of tasks too long for a productive
user experience, and also creates an unnecessary load on the master.
It would be awesome if we could filter tasks by label, which can be customized
by the user, so we could do something like:
`curl /tasks?<label1>=<value1>&<label2>=<value2>`
we could of course also do something fancy like using a resource query
language, but that can be up for debate.
> Allow for custom filters on Mesos APIs
> ---------------------------------------
>
> Key: MESOS-7170
> URL: https://issues.apache.org/jira/browse/MESOS-7170
> Project: Mesos
> Issue Type: Improvement
> Components: HTTP API
> Reporter: Miguel Bernadin
> Assignee: Gilbert Song
> Priority: Minor
>
> For tasks.json API and others like state.json, etc, on larger clusters the
> data that Mesos master sends is quite lengthy. It would be good to provide
> filters in the API to allow Mesos to just send only the RUNNING tasks in the
> cluster so it does less work. Creating this JIRA so we can have intelligent
> filters to pick what data to send on the server side, rather than filtering
> it out on the client side.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
