Add error tracking metrics to API main loop exception handlers

[mgazza]

Summary

The main loop exception handlers in GECloud, GECloudData, Fox, and Octopus API clients silently swallow exceptions - they log them but don't track them for monitoring purposes. This makes it difficult to detect when API clients are experiencing repeated errors.

Problem

When an exception occurs in any of these main loops:

• GECloud (apps/predbat/gecloud.py:875)
• GECloudData (apps/predbat/gecloud.py:1455)
• Fox API (apps/predbat/fox.py:142)
• Octopus API (apps/predbat/octopus.py:440)

The error is logged but:

1. failures_total counter is not incremented (inconsistent with API request failures)
2. had_errors flag is not set (so monitoring systems can't detect issues)

Impact

• Monitoring systems can't alert on repeated main loop exceptions
• The error metrics stay at 0 even when errors are occurring repeatedly
• Makes it hard to detect and diagnose issues like the recent EVC device bug (GECloud: async_get_evc_device missing 'previous' parameter causes TypeError #2931)

Proposed Solution

Add error tracking to each main loop exception handler:

except Exception as e:
    self.log("Error: ... {}".format(e))
    self.failures_total += 1
    if hasattr(self.base, 'had_errors'):
        self.base.had_errors = True

This ensures:

1. API failure metrics are incremented (consistent with other API failure tracking)
2. General error flag is set for monitoring systems that check had_errors
3. Uses hasattr() check for backwards compatibility

[gcoan]

If you are looking at the error handling, there are other issues such as Predbat crashing if it can't retrieve an external API #2043 #2062 #2054