[Doc] added description of monitoring setting write_direct #11586
Conversation
There was a problem hiding this comment.
What you have here builds cleanly and LGTM.
Should this new setting be added here:
Line 228 in 5f50b7f
|
I've added, but in this other PR #11541 |
|
|
||
| Forces the internal collectors to write directly to the monitoring {es} without | ||
| traversing the monitoring exporter provided by {es}. Disabled by default, this setting | ||
| is intermediate step in deprecating the internal collectors. |
There was a problem hiding this comment.
Adding to my comment with the code changes, I suspect we are also going to need to do some more extensive documentation changes for this update to mirror the docs for filebeat monitoring, where there are three options for monitoring: Legacy collection (monitoring data sent via production cluster), Internal collection (monitoring data sent direct to monitoring cluster) and metricbeat collection.
There was a problem hiding this comment.
I've changed a little bit the documenatiton part related to xpack monitoring, let me know what do you think @karenzone
andsel
force-pushed
the
feature/document_write_direct_setting
branch
from
1f0d711
to
bfd7a04
Compare
…xed all the `xpack.monitoring.*` references
There was a problem hiding this comment.
I think the monitoring docs need a thorough rework. The changes are more extensive than just the changes to settings, and I need to update the docs to specify the cluster to send monitoring data, etc.
I'd like to work on this as a separate PR, and get input from you, Rob, others on the structure and content. WDYT?
|
@karenzone If you are willing to do that, that would be awesome. I think this PR is a good jumping off point, but there are a few areas in the (new) internal collection section that could do with a bit of clarification/clean up - the docs still reference sending to the production cluster, and it might be nice to talk about using the new |
|
I agree with you |
Describe the setting
xpack.monitoring.collection.write_direct.enabledintroduced with PR #11541. The behaviour of this flag is to enable the internal collectors to shape the metrics documents in a way to skip the Elasticsearch's monitoring exporter. Essentially we do the same job the exporter already does, and we switch the endpoint we use to push data from_monitoring/bulkto_bulk/monitoring