UI/UX Constant Error Popups Caused by Polling on Failed API Requests

**What is the problem, where is it and why is it an issue?**

When viewing a cluster, users frequently encounter repeated error pop-ups every few seconds due to continuous frontend polling. This results in a poor user experience, as the error dialog keeps reappearing whenever an API endpoint fails or a service is temporarily unavailable, as described below.

Reported Example:

### Example Endpoint:
- nodes/metrics service continuously returns 503 Service Unavailable

### Reproduction Steps (Example)

1. Create a cluster. (e.g OpenStack)
2. Navigate to Cluster Details → Machine Deployment Details (Nodes).
3. Observe that the `nodes/metrics` endpoint fails periodically on each refresh interval.

### Impact
- Results in **poor UX** due to repeated error popups triggered by polling.

### **Suggest how to fix it, if possible**

We should explore ways to improve error handling and suppression in such cases for example:

- Improving the UX to handle repeated API failures more gracefully
- Considering backoff or retry strategies for repeated polling failures.
- Consider switching from polling to websocket instead of polling.

Reproduce the issue from Cluster Details → Machine Deployment Details (Nodes) with the nodes/metrics endpoint returning 503. Trace the polling and error-dialog flow for that view, then evaluate the suggested suppression, backoff, retry, or websocket approaches. Done means repeated polling failures no longer cause recurring popups while the failure remains observable.

Written by the indexing model from the issue text.