Track a 'more' metric for the Updater

[michelle192837]

The Updater logs when it has more work to do after updating a test group (with 'more = true' in the log line), but it'd be nice to have a metric that allows us to track if there are spikes in groups that don't update, or a consistent line of groups that can't catch up on repeated updates. (This is expected in small numbers because some group updates, especially for new groups, take a long time, but it's a bad sign if this climbs or stays high).

[moki1202]

@michelle192837 Can this be solved by making a map and every time the update function returns an error, we log the results in our map. Plus we can include a counter to it.

[michelle192837]

We actually have a standard metrics interface use throughout components that can be used for this! We just need to add that to the list of metrics tracked, and increment at the right time.

More specifically:

• The updater defines the metrics it cares about at https://github.com/GoogleCloudPlatform/testgrid/blob/master/pkg/updater/updater.go#L57. The functions below that are convenience methods to update those metrics as the updater does work. We need a new metric here to count the times the updater hits the case below.
• The updater tracks whether there are additional columns to be updated (more = true), finally updating a log line at https://github.com/GoogleCloudPlatform/testgrid/blob/master/pkg/updater/updater.go#L732. This should also update a metric.

[moki1202]

/assign