Fail squid process with more obvious message when it fails to construct processor instance

[bvotteler]

When squid runs into an exception during the construction of the processor instance (as it did with failures to interpret updated metadata in the past), it quits with an unclear message. Our infrastructure will automatically restart it, so it will fail again in a loop.

When looking at the logs, it wasn’t immediately obvious that the process is in a crash-loop as this happened. The messages could simply be interpreted as some error that is non-fatal, caught and logged.

Therefore the task is:
Catch those exceptions and explicitly log that the process will be terminating before rethrowing (or doing a process.exit() or similar) so it is clear that squid is shutting down.