Last updated: October 3, 2025
Code Red Newsletter #17
Hi there,
I’ve spent the past couple of weeks deep in the weeds on ingress controllers - Ingress-NGINX, Contour, Traefik, Emissary - and how to light them up with OpenTelemetry. The result: a series on observability at the edge and how to make the front door of your platform part of your distributed traces.
In Focus: Observing the Edge of your cluster
Why bother? Because the ingress layer is the real beginning of every user journey. If your ingress is slow, everything feels slow. If it drops requests, users never even reach your services. And without ingress-level spans, your traces are incomplete - you’re only telling half the story.
Observability at the edge gives you end-to-end traces that start at the entry point, correlated logs that explain why requests behave the way they do, and metrics that reveal trends in latency and retries before they ripple downstream. That’s why we kicked off this series: to show how Ingress-NGINX, Contour, Traefik, and Emissary can all be made fully observable with OpenTelemetry and Dash0.
Dash0 Raises $35 Million Series A to Build the First AI-Native Observability Platform
We’ve raised a $35M Series A to double down on our mission of fixing observability. Too many teams still wrestle with siloed signals, proprietary agents, and bills that grow faster than insights.
From day one we’ve taken a different path: open standards like OpenTelemetry and Perses, self-service onboarding, our telemetry Spam Filter, and one-click Triage. With this funding, we’re accelerating the next step - AI-native observability. Agent0, our upcoming SRE AI agent, will help turn observability from insights into action.
Thanks for your support - we couldn’t have reached this milestone without you.
Read the full announcement here.
Observing Ingress-NGINX with OpenTelemetry and Dash0
Ingress-NGINX might be in unofficial maintenance mode, but it still fronts more clusters than anything else. It has solid tracing support, structured logs that can be parsed into spans, and Prometheus metrics that can be enriched with Kubernetes metadata. The maturity is high for traces and metrics, with logs needing a bit more work.
Observing Contour with OpenTelemetry and Dash0
Contour leans on Envoy, and that means first-class tracing with context propagation baked in. Metrics are plentiful (sometimes too plentiful), and logs can be parsed for traceparent headers to correlate with spans. Overall, a strong edge option with traces nearly turnkey and metrics rich but verbose.
Observing Traefik with OpenTelemetry and Dash0
Traefik embraces OpenTelemetry natively. It can export traces, metrics, and even experimental logs directly as OTLP - no scraping required. That makes it one of the cleanest controllers to work with, and one of the most future-proof for full signal correlation.
Oh, and by the way, see below the link to a webinar we had with the awesome people at Traefik about this.
Observing Emissary Ingress with OpenTelemetry and Dash0
Emissary’s Envoy core and Gateway API alignment make it a forward-looking edge option. Traces are well supported, metrics are exposed on a dedicated endpoint, and logs can be enriched with span context. Its OTel maturity sits comfortably between Contour and Traefik.
Observability at the Edge: Comparing Kubernetes Ingress Controllers
To wrap up the series, we published an overview comparing all four ingress controllers: what each supports natively, where Prometheus or parsing is still needed, and how the OpenTelemetry Collector closes the gaps.
Read the comparison post here.
Choice Cuts
You’ve had your main course of ingress deep dives - now here are a few extra servings from around the ecosystem.
GCP: OpenTelemetry now in Google Cloud Observability
Google Cloud now supports OTLP ingest directly in Google Cloud Trace. For GCP users, that means no more proprietary agents - just standard OpenTelemetry pipelines into a managed backend. It’s a great start, and we look forward to the day when every Google Cloud service can also export telemetry natively as OTLP.
Is It Observable: Kubernetes v1.33 Admission Controller Updates
Henrik Rexed breaks down Kubernetes 1.33’s new admission controller. CEL-powered policies now come with better hooks for monitoring and debugging policy violations.
Kill the Bill: Cut Data Costs, Elevate Insights with Traefik & Dash0
Telemetry bills piling up? Michele Mancioppi (Dash0) and Immánuel Fodor (Traefik) walked through how to cut noise while keeping high-fidelity signals, blending OTel best practices with a live demo of precision filtering in action. If you missed it live, you can still catch the full recording.
Final Thoughts
Ingress controllers aren’t just gatekeepers - they’re the start of every trace, the first line of logs, and a rich source of metrics when observed right. With OpenTelemetry, they’re no longer a blind spot but a foundation for end-to-end observability.
And with our Series A, we’re even better positioned to help teams put this into practice: cutting data costs, connecting signals, and turning the edge of your cluster into a source of clarity.
Next week is going to be busy:
- Cloud Native Denmark, Aarhus, October 7–8 - I’m organizing and hosting this two-day community conference. If you’re local, come by for talks, workshops, and a lot of hallway track energy.
- SREcon, Dublin, October 7–9 - find Dash0 at Booth #36. Into site reliability and distributed systems? Stop by for a chat about how OpenTelemetry-native observability can level up your platform.
- AWS Community Day DACH, Munich, October 7 - Expect inspiring talks, networking, and a chance to see Dash0 in action with demos, walkthroughs, and limited-edition backpacks.
Heading to Aarhus, Dublin, or Munich? Don’t miss it - swing by and let’s connect.
See you in two weeks.
Kasper, out.
