Mailing-List: contact dev-help@ignite.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@ignite.apache.org
Date: Wed, 6 Dec 2017 15:23:00 +0000 (UTC)
From: "Denis Mekhanikov (JIRA)" <jira@apache.org>
To: dev@ignite.apache.org
Message-ID: <JIRA.13123217.1512573730000.399175.1512573780194@Atlassian.JIRA>
In-Reply-To: <JIRA.13123217.1512573730000@Atlassian.JIRA>
References: <JIRA.13123217.1512573730000@Atlassian.JIRA> <JIRA.13123217.1512573730052@jira-lw-us.apache.org>
Subject: [jira] [Created] (IGNITE-7125) Cluster metrics for a single node
 differ from local metrics
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
archived-at: Wed, 06 Dec 2017 15:23:10 -0000

Denis Mekhanikov created IGNITE-7125:
----------------------------------------

             Summary: Cluster metrics for a single node differ from local m=
etrics
                 Key: IGNITE-7125
                 URL: https://issues.apache.org/jira/browse/IGNITE-7125
             Project: Ignite
          Issue Type: Bug
    Affects Versions: 2.3
            Reporter: Denis Mekhanikov


Some metrics, that perform aggregation of multiple values, have different m=
eanings for a local node and for a cluster. Local metrics perform aggregati=
on over time, but cluster metrics =E2=80=93 over nodes in the cluster. As a=
 result, you can get different values of local metrics and cluster metrics,=
 when there is only one node in the cluster.

Here is a reproducer:
{code}
public class ExampleNodeStartup {
    public static void main(String[] args) throws Exception {
        Ignite ignite =3D Ignition.start();

        while (true) {
            ClusterMetrics locMetrics =3D ignite.cluster().localNode().metr=
ics();
            ClusterMetrics clusterMetrics =3D ignite.cluster().metrics();

            System.out.println("averageCpuLoad: " +
                "local=3D" + locMetrics.getAverageCpuLoad() +
                "; cluster=3D" + clusterMetrics.getAverageCpuLoad());
            System.out.println("maximumThreadCount: " +
                "local=3D" + locMetrics.getMaximumThreadCount() +
                "; cluster=3D" + clusterMetrics.getMaximumThreadCount());

            Thread.sleep(200);
        }
    }
}
{code}

After some time values begin to differ.

All {{ClusterMetrics#getMaximum*}} and {{ClusterMetrics#getAverage*}} metri=
cs seem to have the same problem.


--
This message was sent by Atlassian JIRA
(v6.4.14#64029)