Add a health check with last time of sync

[linki]

This adds an extensible health check endpoint as well as an initial check for the last sync time.

The output can be seen at http://127.0.0.1:7979/healthz?full=1.

This particular check starts out unhealthy and switches to healthy for as long as the last sync interval isn't missed for too long (this code sets it to two intervals).

This can be used

• in a readinessProbe to ensure the initial sync succeeds, as well as
• in a livenessProbe to restart stackset-controller in case it gets stuck

[coveralls]

Pull Request Test Coverage Report for Build 1480

• 1 of 12 (8.33%) changed or added relevant lines in 1 file are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage decreased (-0.3%) to 69.17%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
controller/stackset.go	1	12	8.33%

Totals
Change from base Build 1464:	-0.3%
Covered Lines:	1575
Relevant Lines:	2277

---

💛 - Coveralls

[aermakov-zalando]

It'll still be healthy before the first Run starts, might want to add a dummy failed check when registering.

[linki]

I moved registration of /healthz to after startWatch() is called. This should avoid any false positives during initial bootstrap.

[aermakov-zalando]

Could you rebase on master? I've fixed the e2e deployment.

[aermakov-zalando]

Otherwise LGTM.

[linki]

👍

[njuettner]

👍