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
Currently if any tunable is changed in yaml where autotune might not understand how to handle it the pod restart fails, so adding them in a stage by stage fashion helps here. Inspired by the process of appsody stacks
I would like to suggest, adding tunable in 3 stages stable, incubator, experimental
Every newly added tunable needs to go from experimental -> incubator -> stable
stable - has everything set
incubator - is having minimal stuff in java to handle , to test and to check
experimental - will be added to yamls but no java code corresponding to that will be there and these ignored by autotune which avoids crashloopbackoff
The text was updated successfully, but these errors were encountered:
Currently if any tunable is changed in yaml where autotune might not understand how to handle it the pod restart fails, so adding them in a stage by stage fashion helps here. Inspired by the process of appsody stacks
I would like to suggest, adding tunable in 3 stages
stable
,incubator
,experimental
Every newly added tunable needs to go from experimental -> incubator -> stable
stable
- has everything setincubator
- is having minimal stuff in java to handle , to test and to checkexperimental
- will be added to yamls but no java code corresponding to that will be there and these ignored by autotune which avoids crashloopbackoffThe text was updated successfully, but these errors were encountered: