[ 
https://issues.apache.org/jira/browse/FELIX-6523?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Ankita Agarwal updated FELIX-6523:
----------------------------------
    Description: 
Felix General HC goes to the Critical state when any of the mandatory bundles 
is not active in BundlesStartedCheck config

*It works in two ways:*

There is a flag "useCriticalForInactive" for bundles configuration in the HC.
 # If the flag "useCriticalForInactive"=true, the HC goes to the Critical state 
when the bundles are not active and then moves back to the OK state if the 
bundle comes back to the active state.
 # If the flag "useCriticalForInactive"=false, the HC goes to the WARN state 
when the bundles are not active and remains in WARN state even after 10mins 
which is not the case with other configurations like DsComponentCheck. Other 
configuration state changes to CRITICAL after 10mins.

 

*Proposed Solution (To make it work same as DsComponentCheck):*

If the flag "useCriticalForInactive"=false, the HC goes to the 
Temporary_Unavailable state when the bundles are not active, and internally we 
will keep on checking the state for 10mins if it doesn’t change to OK then we 
will move it to the Critical state.

If the flag "useCriticalForInactive"=true, the HC goes straight to the Critical 
state when the bundles are not active, and internally we will keep on checking 
the state to move it back OK state.

  was:
Felix General HC goes to the Critical state when any of the mandatory bundles 
is not active in BundlesStartedCheck config

It works in two ways:

There is a flag "useCriticalForInactive" for bundles configuration in the HC.
 # If the flag "useCriticalForInactive"=true, the HC goes to the Critical state 
when the bundles are not active and then moves back to the OK state if the 
bundle comes back to the active state.
 # If the flag "useCriticalForInactive"=false, the HC goes to the WARN state 
when the bundles are not active and remains in WARN state even after 10mins 
which is not the case with other configurations like DsComponentCheck. Other 
configuration state changes to CRITICAL after 10mins.

 

Proposed Solution (To make it work same as DsComponentCheck):

If the flag "useCriticalForInactive"=false, the HC goes to the 
Temporary_Unavailable state when the bundles are not active, and internally we 
will keep on checking the state for 10mins if it doesn’t change to OK then we 
will move it to the Critical state.

If the flag "useCriticalForInactive"=true, the HC goes straight to the Critical 
state when the bundles are not active, and internally we will keep on checking 
the state to move it back OK state.


> BundlesStartedCheck status is not aligned with other configs in the General HC
> ------------------------------------------------------------------------------
>
>                 Key: FELIX-6523
>                 URL: https://issues.apache.org/jira/browse/FELIX-6523
>             Project: Felix
>          Issue Type: Improvement
>          Components: Health Checks
>    Affects Versions: healthcheck.core 2.0.14
>            Reporter: Ankita Agarwal
>            Priority: Major
>
> Felix General HC goes to the Critical state when any of the mandatory bundles 
> is not active in BundlesStartedCheck config
> *It works in two ways:*
> There is a flag "useCriticalForInactive" for bundles configuration in the HC.
>  # If the flag "useCriticalForInactive"=true, the HC goes to the Critical 
> state when the bundles are not active and then moves back to the OK state if 
> the bundle comes back to the active state.
>  # If the flag "useCriticalForInactive"=false, the HC goes to the WARN state 
> when the bundles are not active and remains in WARN state even after 10mins 
> which is not the case with other configurations like DsComponentCheck. Other 
> configuration state changes to CRITICAL after 10mins.
>  
> *Proposed Solution (To make it work same as DsComponentCheck):*
> If the flag "useCriticalForInactive"=false, the HC goes to the 
> Temporary_Unavailable state when the bundles are not active, and internally 
> we will keep on checking the state for 10mins if it doesn’t change to OK then 
> we will move it to the Critical state.
> If the flag "useCriticalForInactive"=true, the HC goes straight to the 
> Critical state when the bundles are not active, and internally we will keep 
> on checking the state to move it back OK state.



--
This message was sent by Atlassian Jira
(v8.20.7#820007)

Reply via email to