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
Using the code added for #281, we should remove static assets from elastic/logs and instead prefer the ones we can get from https://github.com/elastic/package-assets. This should mean removing a bunch of jinja and focusing on dynamic retrieval per the mechanisms that are in place now.
Required packages for parity:
kafka
nginx
apache
mysql
kafka
redis
system
One ambiguous thing to handle is going to be the so-called application-logs integration, which is not an integration at all but seems to be some carryover filebeat/libbeat data. We may want to leave it behind at this point, provided its absence doesn't wildly change the profile of the benchmark.
High level steps:
Onboard target integrations to package-assets workflow for publishing (from productionpackage-storage, as a start)
Change the track to use these integrations, testing for functional badness with our old-ish data but also drastic, unquantifiable performance differences
Follow-up to include generating or gathering more up-to-date seed data for the corpora
The text was updated successfully, but these errors were encountered:
Quick update: I'm chipping away at this with a local copy of package-assets. There are a few gotchas so far, but the overall strategy seems like it's going to work out. I should have a PoC with a subset of the logging track to share shortly.
Using the code added for #281, we should remove static assets from
elastic/logs
and instead prefer the ones we can get from https://github.com/elastic/package-assets. This should mean removing a bunch of jinja and focusing on dynamic retrieval per the mechanisms that are in place now.Required packages for parity:
One ambiguous thing to handle is going to be the so-called
application-logs
integration, which is not an integration at all but seems to be some carryover filebeat/libbeat data. We may want to leave it behind at this point, provided its absence doesn't wildly change the profile of the benchmark.High level steps:
package-assets
workflow for publishing (fromproduction
package-storage
, as a start)Follow-up to include generating or gathering more up-to-date seed data for the corpora
The text was updated successfully, but these errors were encountered: