cli,workload: don't hide useful workloads #93992
Conversation
knz
requested review from
herkolategan and
renatolabs
and removed request for
a team
|
This change is |
|
@j82w I'm nominating you to review this because we don't have any one team responsible for this part of our infrastructure, and your opinion is thus at least as good as anyone else. I even think your opinion is actually better since you've recently created a workload that deserves this treatment. |
knz
force-pushed
the
20221220-workloads
branch
2 times, most recently
from
bbf7fa1 to
ab3826a
Compare
There was a problem hiding this comment.
I agree that making them public could be helpful for users. I don't see an issue with it. Not sure if any of the other teams have a reason for keeping them hidden. The plan was to make insights workload public, but wanted to get feedback on the initial implementation before doing so.
Reviewed 16 of 16 files at r1, all commit messages.
Reviewable status:complete! 0 of 0 LGTMs obtained (waiting on @herkolategan, @knz, and @renatolabs)
pkg/workload/ttllogger/ttllogger.go line 50 at r1 (raw file):
Name: "ttllogger", Description: "Generates a simple log table with rows expiring after the given TTL.", Version: "0.0.1",
Should this be updated to 1.0.0 to match most of the other versions since it will be public? Same thing on ttlbench.go
There was a problem hiding this comment.
TFYR!
bors r=j82w
Reviewable status:
pkg/workload/ttllogger/ttllogger.go line 50 at r1 (raw file):
Previously, j82w wrote…
Should this be updated to 1.0.0 to match most of the other versions since it will be public? Same thing on ttlbench.go
I think we should trust the teams who authored these tests to maintain the version numbers themselves. This particular value doesn't hurt.
|
bors r- |
|
Canceled. |
Prior to this patch, at least half of the workloads were hidden from view in the output of `cockroach --help`. There was no good reason for this: most of the workloads are useful for teaching/learning and for experimentation. They all deserve more exposure, so that folk can learn about them without being told by the one person who built the workload in the first place. So this patch fixes that by exposing of all of them through the online help. One question that could remain is how much teaching value there is in letting someone experiment with a tool that was built for the benefit of one team only. One specific workload is under consideration here: `bulkingest`, used for benchmarking inside the D&R team, does not really do anything akin to what an end-user would possibly expect to do with a database. For that workload, and the benefit of future workloas akin to it, this patch adds a notice in its help text that it was developed for internal testing only. Release note: None
knz
force-pushed
the
20221220-workloads
branch
from
ab3826a to
df0ef24
Compare
|
bors r=j82w |
There was a problem hiding this comment.
Reviewed 1 of 1 files at r2, all commit messages.
Reviewable status:
|
Build succeeded: |
Fixes #94079.
Prior to this patch, at least half of the workloads were hidden from view in the output of
cockroach --help.There was no good reason for this: most of the workloads are useful for teaching/learning and for experimentation. They all deserve more exposure, so that folk can learn about them without being told by the one person who built the workload in the first place.
So this patch fixes that by exposing of all of them through the online help.
One question that could remain is how much teaching value there is in letting someone experiment with a tool that was built for the benefit of one team only. One specific workload is under consideration here:
bulkingest, used for benchmarking inside the D&R team, does not really do anything akin to what an end-user would possibly expect to do with a database. For that workload, and the benefit of future workloas akin to it, this patch adds a notice in its help text that it was developed for internal testing only.Release note: None
Epic: None