fix: raise DeviceNotFound exception if device is not online in filter

[guptadev21]

Description

This PR optimizes the behavior of the get_intf_ip filter when applied to devices that are not online.
Previously, the filter would always invoke device.ready(), which internally polls the device up to 18 times before giving up. This led to unnecessary delays and wasted time in cases where the device was already offline.

With this change, the filter first checks whether the device is online. If the device is not online, it will skip the ready-state polling and fail fast, avoiding the long wait.

Why this change is needed

• Prevents unnecessary polling on offline devices.
• Reduces latency when executing filters over devices that are known to be offline.
• Improves user experience by making filter execution faster and more predictable.

Impact

• Only affects get_intf_ip filter logic.
• Online devices will continue to behave as before.
• Offline devices will now short-circuit without waiting for the 18 polling cycles.

[github-actions]

🤖 Pull Request Artifacts (#18032272877) 🎉

[amitsingh21]

it might not solve the problem, say for example if device is offline, now we poll it ready and bailout with exception and the resource which was dependent on this function will fail resulting into incomplete resources in the system. Instead this flow is more resilent:

• have a function named poll_till_intf_ready() which will with a retry wait for say same period as we are doing currently for 180s until interface is available in devices response
• since the poller on cloud side runs every 120s we are bound to get interface if interface is there on the device.
• in case of registered state also if device is actually onboarded we will get the interface ready within 180s