[JBoss JIRA] (HAWKULARQE-74) 2nd baselines
by viet nguyen (JIRA)
[ https://issues.jboss.org/browse/HAWKULARQE-74?page=com.atlassian.jira.plu... ]
viet nguyen edited comment on HAWKULARQE-74 at 3/27/17 1:26 AM:
----------------------------------------------------------------
I hit a hard limit at 240 pods on the node !239pods.png|thumbnail!
239 test pods + 1 HOSA pod.
{code}
[root@b20 ~]# free -g
total used free shared buff/cache available
Mem: 62 16 32 3 13 41
Swap: 0 0 0
{code}
- HOSA collection interval 30seconds
- Use [Hawkular Python client |https://github.com/vnugent/pyme]and iterate through each pod counting number of raw metrics captured in 30 minute-duration.
- 2 raw metrics / minute * 30 minute = 60 metrics
Observations:
- Python client -> Metrics master endpoint on slow VPN connection causes exceptions in Metrics log [BZ 1435436|https://bugzilla.redhat.com/show_bug.cgi?id=1435436]
was (Author: vietn):
I hit a hard limit at 240 pods on the node !239pods.png|thumbnail!
239 test pods + 1 HOSA pod.
{code}
[root@b20 ~]# free -g
total used free shared buff/cache available
Mem: 62 16 32 3 13 41
Swap: 0 0 0
{code}
- HOSA collection interval 30seconds
- Use [Hawkular Python client |https://github.com/vnugent/pyme]and iterate through each pod counting number of raw metrics captured in 30 minute-duration.
- 2 raw metrics / minute * 30 minute = 60 metrics
> 2nd baselines
> -------------
>
> Key: HAWKULARQE-74
> URL: https://issues.jboss.org/browse/HAWKULARQE-74
> Project: Hawkular QE
> Issue Type: Sub-task
> Reporter: viet nguyen
> Assignee: viet nguyen
> Attachments: 239pods.png, 30minutes-raw.txt.zip
>
>
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)
9 years, 1 month
[JBoss JIRA] (HAWKULARQE-74) 2nd baselines
by viet nguyen (JIRA)
[ https://issues.jboss.org/browse/HAWKULARQE-74?page=com.atlassian.jira.plu... ]
viet nguyen edited comment on HAWKULARQE-74 at 3/27/17 1:24 AM:
----------------------------------------------------------------
I hit a hard limit at 240 pods on the node !239pods.png|thumbnail!
239 test pods + 1 HOSA pod.
{code}
[root@b20 ~]# free -g
total used free shared buff/cache available
Mem: 62 16 32 3 13 41
Swap: 0 0 0
{code}
- HOSA collection interval 30seconds
- Use [Hawkular Python client |https://github.com/vnugent/pyme]and iterate through each pod counting number of raw metrics captured in 30 minute-duration.
- 2 raw metrics / minute * 30 minute = 60 metrics
was (Author: vietn):
I hit a hard limit at 240 pods on the node !239pods.png|thumbnail!
239 test pods + 1 HOSA pod.
{code}
[root@b20 ~]# free -g
total used free shared buff/cache available
Mem: 62 16 32 3 13 41
Swap: 0 0 0
{code}
- HOSA collection interval 30seconds
- Use [Hawkular Python client |https://github.com/vnugent/pyme]and iterate through each pod counting number of raw metrics in 30 minutes in the past.
- 2 custom metrics / pod * 2/minutes * 30 minute = 60 metrics/30minutes
> 2nd baselines
> -------------
>
> Key: HAWKULARQE-74
> URL: https://issues.jboss.org/browse/HAWKULARQE-74
> Project: Hawkular QE
> Issue Type: Sub-task
> Reporter: viet nguyen
> Assignee: viet nguyen
> Attachments: 239pods.png, 30minutes-raw.txt.zip
>
>
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)
9 years, 1 month
[JBoss JIRA] (HAWKULARQE-74) 2nd baselines
by viet nguyen (JIRA)
[ https://issues.jboss.org/browse/HAWKULARQE-74?page=com.atlassian.jira.plu... ]
viet nguyen edited comment on HAWKULARQE-74 at 3/27/17 1:21 AM:
----------------------------------------------------------------
I hit a hard limit at 240 pods on the node !239pods.png|thumbnail!
239 test pods + 1 HOSA pod.
{code}
[root@b20 ~]# free -g
total used free shared buff/cache available
Mem: 62 16 32 3 13 41
Swap: 0 0 0
{code}
- HOSA collection interval 30seconds
- Use [Hawkular Python client |https://github.com/vnugent/pyme]and iterate through each pod counting number of raw metrics in 30 minutes in the past.
- 2 custom metrics / pod * 2/minutes * 30 minute = 60 metrics/30minutes
was (Author: vietn):
I hit a hard limit at 240 pods on the node !239pods.png|thumbnail!
239 test pods + 1 HOSA pod.
{code}
[root@b20 ~]# free -g
total used free shared buff/cache available
Mem: 62 16 32 3 13 41
Swap: 0 0 0
{code}
> 2nd baselines
> -------------
>
> Key: HAWKULARQE-74
> URL: https://issues.jboss.org/browse/HAWKULARQE-74
> Project: Hawkular QE
> Issue Type: Sub-task
> Reporter: viet nguyen
> Assignee: viet nguyen
> Attachments: 239pods.png, 30minutes-raw.txt.zip
>
>
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)
9 years, 1 month
[JBoss JIRA] (HAWKULARQE-74) 2nd baselines
by viet nguyen (JIRA)
[ https://issues.jboss.org/browse/HAWKULARQE-74?page=com.atlassian.jira.plu... ]
viet nguyen edited comment on HAWKULARQE-74 at 3/27/17 1:12 AM:
----------------------------------------------------------------
I hit a hard limit at 240 pods on the node !239pods.png|thumbnail!
239 test pods + 1 HOSA pod.
{code}
[root@b20 ~]# free -g
total used free shared buff/cache available
Mem: 62 16 32 3 13 41
Swap: 0 0 0
{code}
was (Author: vietn):
I hit a hard limit at 239 pods on the node !239pods.png|thumbnail!
{code}
[root@b20 ~]# free -g
total used free shared buff/cache available
Mem: 62 16 32 3 13 41
Swap: 0 0 0
{code}
> 2nd baselines
> -------------
>
> Key: HAWKULARQE-74
> URL: https://issues.jboss.org/browse/HAWKULARQE-74
> Project: Hawkular QE
> Issue Type: Sub-task
> Reporter: viet nguyen
> Assignee: viet nguyen
> Attachments: 239pods.png, 30minutes-raw.txt.zip
>
>
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)
9 years, 1 month
[JBoss JIRA] (HAWKULARQE-73) Rebuild internal cluster 2
by viet nguyen (JIRA)
[ https://issues.jboss.org/browse/HAWKULARQE-73?page=com.atlassian.jira.plu... ]
viet nguyen updated HAWKULARQE-73:
----------------------------------
Description:
Use cluster2 (JON BC hardware) for a larger set up than OS1.
- Rebuild cluster. Both master and node are on bare metal
- 30 metrics per pod
- -how many pods?- 240 pods
- verify custom metrics definition count - 2 per pod
- verify metrics data - 30s interval, 2 per minute, 60 per 30 minutes
Cluster info: OSE 3.5
- Master, infra node 46G RAM
-- metrics, heapster, HOSA, 30s collection internal
- Regular node 64G RAM
-- 239 promgen pods (prometheus generator) + HOSA pod
was:
Use cluster2 (JON BC hardware) for a larger set up than OS1.
- Rebuild cluster. Both master and node are on bare metal
- 30 metrics per pod
- how many pods?
- verify metrics definition count
- verify metrics data
Cluster info: OSE 3.5
- Master, infra node 46G RAM
-- metrics, heapster, HOSA
- Regular node 64G RAM
-- 239 promgen pods (prometheus generator) + HOSA pod
> Rebuild internal cluster 2
> ---------------------------
>
> Key: HAWKULARQE-73
> URL: https://issues.jboss.org/browse/HAWKULARQE-73
> Project: Hawkular QE
> Issue Type: Sub-task
> Reporter: viet nguyen
> Assignee: viet nguyen
>
> Use cluster2 (JON BC hardware) for a larger set up than OS1.
> - Rebuild cluster. Both master and node are on bare metal
> - 30 metrics per pod
> - -how many pods?- 240 pods
> - verify custom metrics definition count - 2 per pod
> - verify metrics data - 30s interval, 2 per minute, 60 per 30 minutes
> Cluster info: OSE 3.5
> - Master, infra node 46G RAM
> -- metrics, heapster, HOSA, 30s collection internal
> - Regular node 64G RAM
> -- 239 promgen pods (prometheus generator) + HOSA pod
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)
9 years, 1 month
[JBoss JIRA] (HAWKULARQE-73) Rebuild internal cluster 2
by viet nguyen (JIRA)
[ https://issues.jboss.org/browse/HAWKULARQE-73?page=com.atlassian.jira.plu... ]
viet nguyen updated HAWKULARQE-73:
----------------------------------
Description:
Use cluster2 (JON BC hardware) for a larger set up than OS1.
- Rebuild cluster. Both master and node are on bare metal
- 30 metrics per pod
- how many pods?
- verify metrics definition count
- verify metrics data
Cluster info: OSE 3.5
- Master, infra node 46G RAM
-- metrics, heapster, HOSA
- Regular node 64G RAM
-- 239 promgen pods (prometheus generator) + HOSA pod
was:
Use cluster2 (JON BC hardware) for a larger set up than OS1.
- Rebuild cluster. Both master and node are on bare metal
- 30 metrics per pod
- -how many pods- 239 promgen pods (prometheus generator)
- verify metrics definition count
- verify metrics data
Cluster info: OSE 3.5
- Master
- -
> Rebuild internal cluster 2
> ---------------------------
>
> Key: HAWKULARQE-73
> URL: https://issues.jboss.org/browse/HAWKULARQE-73
> Project: Hawkular QE
> Issue Type: Sub-task
> Reporter: viet nguyen
> Assignee: viet nguyen
>
> Use cluster2 (JON BC hardware) for a larger set up than OS1.
> - Rebuild cluster. Both master and node are on bare metal
> - 30 metrics per pod
> - how many pods?
> - verify metrics definition count
> - verify metrics data
> Cluster info: OSE 3.5
> - Master, infra node 46G RAM
> -- metrics, heapster, HOSA
> - Regular node 64G RAM
> -- 239 promgen pods (prometheus generator) + HOSA pod
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)
9 years, 1 month
[JBoss JIRA] (HAWKULARQE-73) Rebuild internal cluster 2
by viet nguyen (JIRA)
[ https://issues.jboss.org/browse/HAWKULARQE-73?page=com.atlassian.jira.plu... ]
viet nguyen updated HAWKULARQE-73:
----------------------------------
Description:
Use cluster2 (JON BC hardware) for a larger set up than OS1.
- Rebuild cluster. Both master and node are on bare metal
- 30 metrics per pod
- -how many pods- 239 promgen pods (prometheus generator)
- verify metrics definition count
- verify metrics data
Cluster info: OSE 3.5
- Master
- -
was:
Use cluster2 (JON BC hardware) for a larger set up than OS1.
- Rebuild cluster. Both master and node are on bare metal
- 30 metrics per pod
- -how many pods- 239 promgen pods (prometheus generator)
- verify metrics definition count
- verify metrics data
> Rebuild internal cluster 2
> ---------------------------
>
> Key: HAWKULARQE-73
> URL: https://issues.jboss.org/browse/HAWKULARQE-73
> Project: Hawkular QE
> Issue Type: Sub-task
> Reporter: viet nguyen
> Assignee: viet nguyen
>
> Use cluster2 (JON BC hardware) for a larger set up than OS1.
> - Rebuild cluster. Both master and node are on bare metal
> - 30 metrics per pod
> - -how many pods- 239 promgen pods (prometheus generator)
> - verify metrics definition count
> - verify metrics data
> Cluster info: OSE 3.5
> - Master
> - -
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)
9 years, 1 month