Report failed installer downloads to fleet

[dantecatalfamo]

💥  Actual behavior

If a host is having trouble downloading an installer due to a poor connection, the installer being too large, no disk space, etc., fleetd will keep retrying the download indefinitely without any issues being indicated on the server. This can lead to a software installation appearing to get "stuck" at installing in the device's upcoming activities.

🧑‍💻  Steps to reproduce

1. Try to install software (not VPP) on a host that has one of the listed issues (installer being too large, no disk space, etc.)
2. Wait

🛠️ To fix

Report the number of download retries or issues encountered before launching the software installer back to the server.

After a certain number of download retries or a timeout is reached, update the installation as failed with details about what happened.

There are two ways we could do this.

1. The failures are reported to the server each time, and state in the db is updated, like a retries or errors column on the host_software_installs table. After retrying a certain number of times, mark the install as failed
2. Orbit tracks the failures internally, and reports the failure in the same way it would report an installer failure.

What is the expected workflow as a result of your proposal?

1. Try to install software on a host
2. fleetd tries to download the installer and fails, reports the failure to the server
3. The installation failure appears in the activities timeline similar to how a regular installation failure would appear
4. The user can see why the software failed and try to fix the host

[noahtalerman]

Hey @dantecatalfamo this seems like it's a bug? We don't want software to get stuck in pending.

Can you please update this issue to bug format?

I moved it over to the #g-mdm release board.

[dantecatalfamo]

Hey @noahtalerman, sure thing