Allow users to check off recommended actions they've done.

[vpetersson]

We need to discuss the actual details of this, but the idea is that users should be able to check off Recommended Actions that they have done.

We will then 'snooze' this item from the dashboard for X hours and show it back again if we detect it again.

[a-martynovich]

Here's my idea.

For each recommended action we need to make a model (subclassing a common RecommendedAction model perhaps). So we'll have a model for "Permissive firewall detected", a model for "Vulnerable packages detected", etc. Each one will have different fields, for example the VulnerablePackageRecommendedAction can have a list of vulnerable packages.

Now when we compile a list of recommended actions, we do it as it is done now, but we first get objects for every recommended action and fill them with data. The message of a recommended action should not be generated at this stage.

The list of recommended actions should ideally be compiled when device information changes, which is either after ping or after a change made by UI user (change of firewall policy for example). We can make it a simple Celery task which runs right after any change.

And when displaying recommended actions, every RecommendedAction should be able to generate its message based on what data it has. And that's where the "snooze" option comes in. If an action is "snoozed" it is not displayed.

There're some caveats, though. For example if a list of vulnerable packages has changed (like another one appeared) and a recommended action for vulnerable packages had been snoozed before, should it re-appear? But we shouldn't worry about it because snoozing is active for a period of one ping.

[vpetersson]

For example if a list of vulnerable packages has changed (like another one appeared) and a recommended action for vulnerable packages had been snoozed before, should it re-appear?

Yeah i think it's fine if it re-appears.