What Are the Best Datadog Alternatives for a 10-Person Startup?

I've been looking at observability tools recently, and one question keeps coming up:

Is Datadog still the best option for a small engineering team, or are there better open-source alternatives?

Datadog is feature-rich, but many startups find that costs grow quickly as they collect more logs, metrics, and traces. That has pushed many teams toward OpenTelemetry-based, self-hosted solutions that reduce vendor lock-in and offer more predictable costs.

Here are the options I'm considering:

• SigNoz – Probably the closest open-source alternative to Datadog. It provides logs, metrics, and traces in one platform with native OpenTelemetry support.
• Grafana + Prometheus + Loki + Tempo – Extremely flexible and widely adopted, but you're responsible for operating several components yourself.
• OpenObserve – A newer all-in-one observability platform focused on efficient storage and OpenTelemetry compatibility.
• Jaeger – Great if distributed tracing is your primary requirement, though you'll still need separate solutions for metrics and logs.

Personally, if I were starting a new company with around 10 engineers, I'd instrument everything with OpenTelemetry first so switching observability backends later is straightforward. The backend can change; your instrumentation shouldn't have to.

What are you using today?

• Still on Datadog?
• Switched to SigNoz or Grafana?
• Running another open-source stack that's worked well?

I'd be interested in hearing what has held up in production and whether the operational overhead of self-hosting has been worth the savings.