[
https://issues.apache.org/jira/browse/IGNITE-21227?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Roman Puchkovskiy updated IGNITE-21227:
---------------------------------------
Description:
If a netty thread or an inbound thread gets blocked, this might cause a node
freeze. We need to monitor such situations to:
# Have ability to see such blockages in logs
# See the thread dump corresponding to the blockage moment to make it possible
diagnose the root cause of the blockage
There is a mechanism in AI2 that allows doing so, and it's (at least partially)
ported to AI3, it is about Ignitethreads and IgniteWorkers.
Netty and inbound threads are currently not using the IgniteWorkers mechanism.
We need to either use it or adapt them to include them in this monitoring.
> Monitor network critical threads for being blocked
> --------------------------------------------------
>
> Key: IGNITE-21227
> URL: https://issues.apache.org/jira/browse/IGNITE-21227
> Project: Ignite
> Issue Type: Improvement
> Reporter: Roman Puchkovskiy
> Priority: Major
> Labels: ignite-3
> Fix For: 3.0.0-beta2
>
>
> If a netty thread or an inbound thread gets blocked, this might cause a node
> freeze. We need to monitor such situations to:
> # Have ability to see such blockages in logs
> # See the thread dump corresponding to the blockage moment to make it
> possible diagnose the root cause of the blockage
> There is a mechanism in AI2 that allows doing so, and it's (at least
> partially) ported to AI3, it is about Ignitethreads and IgniteWorkers.
> Netty and inbound threads are currently not using the IgniteWorkers
> mechanism. We need to either use it or adapt them to include them in this
> monitoring.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
