[surfacers.stackdriver] Use random task ID for Google Cloud Run. (clo…

…udprober#267)

See cloudprober#257 for the background.

Cloud Run starts/restarts the container multiple with the same revision id. This creates a problem for the Stackdriver metrics (at least in the short term) as for the cumulative metrics, a point's startTime is tied to the Cloudprober start time and its endTime cannot be before the startTime. This condition is violated if there are multiple processes exporting data to the same resource.

[surfacer.stackdriver] Ignore if we are not able to get the instance …

…info. (cloudprober#228)

Migrate ingresses from v1beta1 api to v1

[cloudprober.http] Export ssl_earliest_cert_expiry_sec as a GAUGE eve…

…ntmetric.

This is to fix: cloudprober#192

[logger] Don't add instance name label on Kubernetes.

[surfacers] Add type infer logic for probestatus. (cloudprober#156)

* [surfacers] Add type infer logic for probestatus.

= Improve inferType tests and make them figure out if type infer logic is missing for some surfacer types.

[probestatus] Reduce point size is there are too many points.

Fix a bug in GCE OAuth error handling. (cloudprober#103)

Fixes cloudprober#102.

Introduce the concept of negative tests (cloudprober#98).

= Add a new probe level option: negative_test.

= If this option is set, mark probe as successful if it doesn't succeed.

= It's currently implemented only for PING and TCP probe types. We return an erorr if this option is set for other probe types.

Fix a minor bug in the contrib Dockerfile.