Observability Without the $50K Datadog Bill

It starts with an invoice. You’ve been on Datadog for a few months, your team is growing, and you open the billing page to find a number that makes you do a double-take. $12,000. Then $28,000. Then you add a few more hosts and enable APM and suddenly you’re staring down $50,000 a year — for a startup that hasn’t hit $1M ARR.

Datadog’s pricing model was designed for enterprises with large infrastructure teams and cost-insensitive procurement processes. Per-host fees. Per-GB ingestion charges. APM as a separate SKU. Custom metrics billed by the thousand. It adds up fast, and it adds up in ways that are hard to predict.

What SaaS Apps Actually Need

Here’s the thing: most SaaS applications don’t need Datadog-scale observability. They need a specific, smaller set of capabilities:

• Traces on critical paths — auth flows, billing webhooks, feature flag evaluations, config changes. Not every database query your ORM generates.
• Tenant-scoped dashboards — when a customer reports an issue, you want to filter everything by their organization instantly. Not grep through logs hoping you tagged things correctly.
• Anomaly alerts — know when a tenant’s error rate spikes, when billing webhooks start failing, when a flag rollout goes sideways.

Datadog can do all of this. But you’ll pay for a lot of things you don’t need to get there, and you’ll spend weeks wiring up the instrumentation correctly.

Flux’s Approach: Auto-Instrumented, Zero Config

Every Flux operation is automatically instrumented. When Flux evaluates a feature flag, processes a billing event, or makes an auth decision, a structured trace is emitted — with full context. No SDK calls. No manual span creation. No trace propagation boilerplate.

This works because Flux owns the operations. We’re not a passive observer layered on top of your code. When we evaluate a flag or check a rate limit, we know exactly what happened, why, and for whom.

Tenant-Aware by Default

Flux traces are scoped to organizations out of the box. Filter your dashboard to a single tenant and see every auth event, flag evaluation, billing operation, and error they’ve experienced — in order, with full context.

This isn’t a nice-to-have. When an enterprise customer files a support ticket, the first thing you want to do is pull up their activity. With Flux, that’s a single filter. Without it, you’re correlating user IDs across five different tools.

OpenTelemetry Export: Not Locked In

If you already have an observability stack — Grafana, Honeycomb, your own Jaeger instance — Flux can export to it via OpenTelemetry. We emit standard OTLP traces and metrics. You can send them to us, to your existing backend, or to both.

We’re not trying to replace your entire observability stack. We’re handling the instrumentation of Flux operations so you don’t have to.

What $500/mo Gets You

The gap widens as you scale. Flux’s pricing is tied to your usage of the platform — not to your infrastructure footprint.

If you’re spending more on observability than on the infrastructure you’re observing, something is wrong. Get started with Flux and close that tab with the Datadog invoice.