Add a second job that watches rekor for new STH entries and publishes

them to GCS.

We don't always have a new one, and under load there might be "batches" of entries
all integrated into the same STH. This means there is no guaranteed frequency of updates
or even a guarantee that every index will exist.

The values (and timestamps) should be monotonically increasing though.

Signed-off-by: Dan Lorenc <dlorenc@google.com>