"Automatic recognition of CIs" stopped working and / or randomly working in SMAX 2023.05

Hi

We have SMAX 2023.05 running on premise version and noticed that autodetection of CIs in Request module has stopped working for some CIs.Such CIs were old existing CIs (not in ended phase).

Such behaviour was seen in the past as well and we can see the SMAX becomes unavailable after a few days of seeing such symptom. Looking at the product documentation about this feature, there's a below statement : 

"The internal database/cache used for CI detection is updated every 5-10 minutes, delays can occur before your device edits are reflected in the detection process."

Does anyone has more details on what this internal database/ cache is, how it gets populated and do we have any control on rebuilding this cache, how it can be monitored for any corruption etc.

The 'global search' on the same CIs are working OK so smart searching is OK.


Thanks

Shishir.

Tags:

  • 0

    Do you see any errors in browser console(developer tools)? or in the backend logs?

  • 0   in reply to 

    There are no errors seen in the browser console.

    Looking at the log file, there's no entry can be found in below logs for the CI being search.

    Issue is that when we paste a bunch of asset tags, only a few asset tags are auto regognised when other asset tags still exists, in NOT ended phase and not protected by data domain. 

    And this issue is quite intermittent thus giving a feeling to user that CI doesn't exists when it actually exists. Once user figure out that CI actually exists, it's losing trust in the functionality.

    In the past, if we restart whole smax and ucmdb, such issue resolves by itself. But we need to know in more detail this time what is causing this problem and if there are small subset of pods that can be restarted without causing the full system wide outage.

    If this feature is dependent on itom-xruntime-platform-offline and cms_gateway pods, can we rollout restart these pods without causing any outage on the overall smax functionality?

    Would be good to know which microservice this feature uses so we only touch those services.

    Thanks

    Shishir.

  • 0 in reply to   

    Better to open a support case to ask that. Otherwise unfortunately you will need to reverse engineer(trace) the functionality i.e like browser console check scripts running when you are in description field... check the backend logs for same CI tag etc.

    It should be a combination of something on change in target field and then querying the backend to find if it matches a CI attribute(ID or whatever is the query field)...

    If you feel that feature is not working as described, then a support case would be a better way.

  • 0   in reply to 

    thanks

    Raising a support case is definitely an option. It was more of an attempt to find out if someone knows which microservice is responsible for providing the CI autodetection feature which we can quickly restart to get back to business as this feature is heavily used by customer for HAM uses case and avoid a full SMAX restart which has seemingly resolved the issue for few days before it reappears.

    BTW, after restarting itom-xruntime-redis pod, I can see CI autodetection is 'improved' and able to find more CIs than before but it is still not finding all the CIs as it should.