-
Notifications
You must be signed in to change notification settings - Fork 29
In-code kubectl apply #767
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
base: master
Are you sure you want to change the base?
Conversation
0eb8125
to
0a5c98f
Compare
20f40d1
to
f2cd750
Compare
f2cd750
to
d124d76
Compare
d124d76
to
abed1e2
Compare
abed1e2
to
d8c7595
Compare
31b264b
to
ec9ed09
Compare
99333c1
to
982953b
Compare
77377f8
to
b0bd34f
Compare
Does this require us to run any tests? I think it's fine to rollout this change and observe after a while, to check this concurrency requirement. Otherwise, PR looks good 👍 |
b0bd34f
to
cd1576f
Compare
Yes, this we can best observe after rolling it out and enabling it in some clusters. |
👍 |
cd1576f
to
d0a458b
Compare
d0a458b
to
0225b7d
Compare
0225b7d
to
41e91bc
Compare
41e91bc
to
fcfc072
Compare
fcfc072
to
1e49116
Compare
1e49116
to
f17a821
Compare
👍 |
Signed-off-by: Mikkel Oscar Lyderik Larsen <[email protected]>
f17a821
to
abce16b
Compare
Proof of Concept of porting the "kubectl apply" code from deployment-service to also utilize it in CLM. Basically this means making the
kubectl apply
operation via Go instead of calling the externalkubectl
binary.The motivation for this change is that we currently restrict how many executions of
kubectl apply
we do concurrently as it's hard to control the memory. This has the side effect of limiting how many clusters we can process in parallel. The idea is that by having this in go we save memory by not having to run another process and we thus can do more in parallel.The feature is behind a config-item
kubectl_apply_inline: true
which enables us to configure it only in a few clusters as a start to validate that it behaves as expected and gives the intended benefits.TODO
Check if we still need to control concurrency.