Meeting Notes from Unomi Community Meeting, November 14th 2024

[Serge Huber <sh...@apache.org>]

Please find below the notes from the Zoom community meeting that occurred
yesterday. Please let me know if I missed anything or if I made any
mistakes.

Regards,

  Serge…


Meeting Notes from Unomi Community Meeting, November 14th 2024

Time: 17:00 CEST


Participants:

   - Serge H. - Inoyu
   - Jonathan S. - Jahia
   - Jérome B. - Jahia
   - Romain G. - Jahia


Minutes :

   - Trouble starting the meeting because the transferred credentials
   didn’t allow to start the meeting. Fortunately François G. came just to
   start it.
   - Serge saw problems when wanting to use debug mode with Docker images,
   it is currently not possible. Will create ticket to make it possible to
   activate debug mode on Docker images using an environment variable.
   Also saw issues with the Groovy actions document about the annotations
   that seems to have been corrected for version 2.6.0, will check.
   Work on OpenSearch continued but not yet completed.
   - Started discussing about PR reviews. The problem is that some PRs stay
   open too long, especially when they are minor changes such as website
   (minor) updates or typos. The idea is to keep using PRs for contributions,
   but to adapt the requirements for reviews.
   Proposal to ping people to ask for reviews or to have a good test
   coverage for lazy reviews.
   We could reviews bypass for minor changes such as minor websites updates
   (date corrections, DOAP file).
   For large PRs the initiator could organize a meeting to present the
   changes, if nobody participates we go to lazy review. Serge will prepare an
   email in the mailing list to propose the changes
   - Release 2.6.0 includes Groovy action documentation updates, migration
   updates from V1 due to volume of data. There were corrections for issues
   with timeouts on calls to Elastic Cloud for long running processes had to
   modify reindex and updatebyquery methods.
   Migration is still very long and it involves a downtime that can be
   quite long.
   For the moment migration has to be done incrementally between versions,
   might be better to be able to go have a system to avoid having to do each
   one sequentially ?
   An idea is to use ES snapshots and a custom journals of ongoing activity
   and only have downtime during journal processing. The process could be
   repeated multiple times if we want to progress the journal « catch up »
   data and do that until the window of missing data is empty.
   It is an interesting project but would only be started if there is a
   client need if the downtimes due to migrations are not acceptable. Serge
   suggested maybe creating a ticket with the idea anyway as it is interesting.
   This would also involve a migration proxy.
   Also included in 2.6.0 are some fixes for problems with special
   characters. A health check was also added to version 2.6.0. Jérome tried to
   use one that exists in Apache Felix but it doesn’t work with our version of
   Karaf.
   - Discussion about updating Karaf: Jérome would like to try to see what
   it involves. Serge suggested to create a branch and start a discussion in
   the mailing list about the blocking issues. We should work on this
   collaboratively, this way is something is stuck others can help.
   - Question about storing completed migrations metadata in ElasticSearch,
   such as the scripts that have completed successfully. Serge agreed on it,
   suggesting it would be a good idea to also include the current version of
   the data model so that we can check if we need to migrate or not. He
   suggested creating a ticket.