[
https://issues.apache.org/jira/browse/IGNITE-28398?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Denis Chudov updated IGNITE-28398:
----------------------------------
Issue Type: Improvement (was: Bug)
> Add meta storage metrics - size of notification futures chain, time of
> notification application
> -----------------------------------------------------------------------------------------------
>
> Key: IGNITE-28398
> URL: https://issues.apache.org/jira/browse/IGNITE-28398
> Project: Ignite
> Issue Type: Improvement
> Reporter: Denis Chudov
> Priority: Major
> Labels: ignite-3
>
> *Motivation*
> Sometimes we see the following:
> {code:java}
> Lease update invocation failed because of outdated lease data on this
> node{code}
> This is logged continuously during the significant period of time.
> This means that meta storage notifications highly likely lag behind the real
> time, which may also affect cluster safe time and maybe something else.
> I think, this may happen because of big chain of notification futures, and in
> the case when each notification takes long time, but now we don't have any
> metrics to see this.
>
> *Definition of done*
> Metrics for size of notification futures chain, time of notification
> application added.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
