ignite-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Nikolay Izhikov <nizhi...@apache.org>
Subject Re: Performance drop with New Monitoring.
Date Mon, 01 Jul 2019 12:17:40 GMT
Hello, Igniter.

I think we should fix the issue with slight changes in internal API (was not released yet).

Changes in API was independently proposed by me and Alexey Plekhanov [1]:

1. We should introduce `MetricGroup` - collection of metric of one specific Ignite entity(io
subsystem, data region, cache, etc.).

2. We can remove entire `MetricGroup` (on cache or cache group destroy) with complexity of
O(1).

3. We can remove metric from `MetricGroup` with complexity of O(1).

[1] https://issues.apache.org/jira/browse/IGNITE-11920?jql=labels%20%3D%20IEP-35%20ORDER%20BY%20status%20ASC

В Пт, 28/06/2019 в 19:10 +0300, Nikolay Izhikov пишет:
> Hello, Maksim.
> 
> I will take a look shortly.
> 
> В Пт, 28/06/2019 в 19:02 +0300, Maksim Stepachev пишет:
> > Hello, Igniters.
> > 
> > I see that the ticket - "[IGNITE-11848] New Monitoring. Phase1" breaks cache stop
performance. It affects next tests: testParallelStartAndStop, testStartManyCaches 
> > 
> > The spent time was increased 6 times:( 
> > 
> > Look at: https://ci.ignite.apache.org/project.html?tab=testDetails&branch_IgniteTests24Java8=%3Cdefault%3E&projectId=IgniteTests24Java8&testNameId=-145777872614811891&page=4

Mime
View raw message