[RFE] Allow max_retries override at instance level

[manageiq]

[RFE] Allow max_retries override at instance level

@mkanoor @tinaafitz What do you think about this ? The idea comes from my implementation on V2V state machines, where the state machine class is just a scaffold. So, I had to set a max_retries for all states as I don't know in advance what it will be used for.

Regarding the impact of such change, I expect it to be minimal as the existing classes and instances wouldn't have to be modified. This adds a new opportunity on top of the existing logic.

[mkanoor]

@fdupont-redhat @tinaafitz @gmcculloug
The max_retries can be set at the instance level, the UI currently doesn't support it, we can expose it in the UI.

Awesome, I didn't expect less from you 👍
Sould I open an RFE in manageiq-ui-classic repo ?

[mkanoor]

Yes, lets open an issue in the UI repo, the MiqAeValue table allows for setting of the max_retries. If they can make these fields show up only for state machines it would be good. e.g. on_entry/on_exit/on_error/max_retries/max_time should only be visible if the field is a state. For relationships and attributes it doesn't make sense.

The Automate Engine might work I am not sure but once the MiqAeValue has the max_retries/max_time we can fix the engine if it doesn't take care of it already.

@mkanoor RFE opened: ManageIQ/manageiq-ui-classic/issues/3560.
Do you know who could look at it ?

[gmcculloug]

I added an assignee to the ui-classic issue.

[tinaafitz]

I manually changed a state machine instance max_retries value and verified the instance value is used in resolution.

[gmcculloug]

@fdupont-redhat With PR ManageIQ/manageiq-ui-classic#3562 merged can this issue and ManageIQ/manageiq-ui-classic#3560 be closed now?

@gmcculloug Yes, we can.