pkg/services/otelhealth: fix uptime_seconds metric type from Gauge to Counter

[pkcll]

Summary

• Change uptime_seconds OTel metric from Float64Gauge to Float64Counter and use Add() instead of Record()
• The Prometheus implementation (promhealth) already correctly uses a Counter — this aligns the OTel implementation to match
• As a Gauge, the metric was recording the fixed interval duration (15s) on every tick instead of accumulating total uptime

[github-actions]

✅ API Diff Results - No breaking changes

---

📄 View full apidiff report

[Copilot]

Pull request overview

Aligns the OpenTelemetry health-check uptime metric with the existing Prometheus implementation by reporting total accumulated uptime rather than repeatedly recording the fixed polling interval.

Changes:

• Switch uptime_seconds from an OTel Float64Gauge to Float64Counter.
• Update reporting from Record() to Add() so uptime accumulates over time.
• Clarify the metric description to indicate it is measured in seconds.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

uptime_seconds is now a Counter and the description says "Uptime of the service in seconds", but it doesn’t indicate the value is cumulative/monotonic (i.e., total uptime accumulated since start). Consider updating the description to explicitly reflect counter semantics to avoid confusion in dashboards/alerts.

Also, this metric represents seconds; most other OTel duration metrics in this repo set an explicit unit (e.g. metric.WithUnit("s") in pkg/settings/limits/time.go). Consider adding metric.WithUnit("s") here as well for consistency and better backend rendering.

Suggested change

	uptimeSeconds, err := meter.Float64Counter("uptime_seconds", metric.WithDescription("Uptime of the service in seconds"))
	uptimeSeconds, err := meter.Float64Counter(
	"uptime_seconds",
	metric.WithDescription("Total cumulative uptime of the service in seconds since start"),
	metric.WithUnit("s"),
	)