Fix #958: add more logging fields to retry job exceptions

[leplatrem]

Fix #958

[grahamalama]

I added some questions about metrics. Maybe instead of including the metrics in each log line, we can add like a "retry run id" (with a better name), and then include that in each log line. Then we can quickly find all of the logs for a particular retry run (with the metrics appearing on the last line as a summary for that run)

[alexcottner]

I'd like to see something added to execute_or_queue in runner.py as well. If execute_action throws an error, we should log an error with the event identifier so we can alert.