You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Describe the enhancement: In addition to upgrading agents when an update is available, allow for downgrade, within reasonable technical restrictions, to an earlier version of Elastic Agent.
Describe a specific use case for the enhancement or feature: The metricbeat memory leak is one example. The workarounds pending that fix were to either disable affected modules (and the scope was unclear enough that I disabled System Metrics but it took another week or two to see I needed to also do it in the Agent integration) or manually downgrading. While some customers may have the capacity to do this with a GPO, Ansible, etc, for others this may be doing this manually or other less ideal methods that sort of defeat one point of centrally managing once the initial deployment is done. A downgrade option would be great for this or other bugs/breaking changes.
Another use case, but maybe not applicable depending on what the root cause was, is the instance or few where a specific version fails to upgrade:
Perhaps this also means downgrades wouldn't work but if not, folks who've upgraded can then downgrade until a fixed version is released without the other workarounds.
The text was updated successfully, but these errors were encountered:
Thanks @jvalente-salemstate for creating this. Anything that is Fleet UI related has to be created in the Kibana repository, however I feel like this issue is probably a duplicate of elastic/kibana#172745.
If we think that's right, let's close this one then.
Describe the enhancement: In addition to upgrading agents when an update is available, allow for downgrade, within reasonable technical restrictions, to an earlier version of Elastic Agent.
Describe a specific use case for the enhancement or feature: The metricbeat memory leak is one example. The workarounds pending that fix were to either disable affected modules (and the scope was unclear enough that I disabled System Metrics but it took another week or two to see I needed to also do it in the Agent integration) or manually downgrading. While some customers may have the capacity to do this with a GPO, Ansible, etc, for others this may be doing this manually or other less ideal methods that sort of defeat one point of centrally managing once the initial deployment is done. A downgrade option would be great for this or other bugs/breaking changes.
Another use case, but maybe not applicable depending on what the root cause was, is the instance or few where a specific version fails to upgrade:
https://github.com/elastic/fleet-server/issues/2383
Perhaps this also means downgrades wouldn't work but if not, folks who've upgraded can then downgrade until a fixed version is released without the other workarounds.
The text was updated successfully, but these errors were encountered: