Configure alerting for unsuccessful invocations

Currently we have alerting configured to detect an absence of successful invocations of the scheduled script, which is a sensible approach given that this will catch even a failure at the GCP level of cloud scheduler failing to trigger. However due to limitations within the alerting period supported by GCP, we can only use this to monitor scripts which are invoked at least once per 24 hours. To catch failures in scripts which are invoked less regularly we can configure explicit failure alerts, which raise based on a non-successful invocation of the cloud function.