-
Notifications
You must be signed in to change notification settings - Fork 8.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[Monitoring] Introduce deprecation warning for shipping monitoring data to the production cluster #51718
Comments
Ping @igoristic. Could you please see if this is something you could take on? Thanks! |
@cachedout @chrisronline Which config settings should trigger the depreciation log? Is it just |
@igoristic None of the existing settings are deprecated. What we need to add is something telling users to switch to using these new configs or using Metricbeat as the collector and shipper of monitoring data. |
@chrisronline For some reason I though that "Internal Monitoring" along with all their config settings are getting deprecated by 8.0 So, if none of the existing "Internal Monitoring" settings are actuating getting deprecated maybe then "depreciation" is a wrong term here? Maybe this should be more of a warning/suggestion instead? I think it might make it kind of confusing for the customer if they are seeing the depreciation warning even-though they are already using Metricbeat. They might think that they still need to change/update something etc. (unless we word it as a general "If you're using..."). So, maybe just form UX point of view it's worth implementing a "smart" detection to see if they are already using Metricbeat to not show anything, and only show the depreciation warning if we detect something like My motivation for this approach is based on: https://github.com/elastic/kibana/blob/master/x-pack/legacy/plugins/monitoring/deprecations.js#L26 |
Yea, it's tricky. The "deprecated settings" in this case is really the configured exporters at the ES level. Kibana only knows to send monitoring data to I think you're right - we might need to take a different approach with this one. We only really care about I don't know if we can use the typical deprecated channels, but we should try because those are where users are used to looking. We need to give users a chance of ensuring a smooth upgrade path in the 7.x timeline |
This is exactly what I need. Thank you Chris! We can include this where we have our current depreciation in the code, and also (like you mentioned) other streams (docs/release-notes) |
@chrisronline I'm struggling to figure out how we can "truly" detect the use of "Internal Monitoring". My original approach of looking for a For my upgrading environment I kept the ES's |
We do something similar with the migration wizard where we look back a limited amount of time, 30 seconds to be specific. If we see both
|
Ah cool! Thanks! Don't know how I missed that. I actually ended up doing something similar, but instead, I'm just making an empty
This might not work so well if they have their collection interval set to something more than This brings up a bigger concerns though, since they can still go back in time and eventually get documents from those none I think it makes more sense to just I might be overthinking this though or describing edge cases (that never happen) 🤷 |
This should be manageable since we won't make schema changes in a minor, and when we go to 8.0, we will provision a new index
But they shouldn't have too. The data in each index should be identical (and we have the parity tests to ensure this)
This is definitely true and we should file a bug to handle this. |
This issue is a request for Kibana to add a deprecation log entry in 7.6 for users who are shipping monitoring data to the production cluster which urges them to switch to using Metricbeat monitoring in order to be prepared for 8.0.
@igoristic is this something you could take on?
cc: @elastic/stack-monitoring
The text was updated successfully, but these errors were encountered: