Customer value: Stop paying Datadog to store metrics you had to drop — move to Elastic Serverless with full coverage at a lower bill.
Datadog bills custom metrics aggressively, so teams drop series, sample metrics, and exclude environments just to keep the invoice down — then miss production blind spots. On the TCO worksheet, Datadog is ~$12.6K/mo (metrics alone ~$3.8K) vs Elastic Serverless ~$1.3K (metrics ~$0.9K). This lab moves the backend: audit what you are not sending to Datadog anymore, then Drop noise at ingest on Elastic instead of paying to store it.
High-cardinality tags and debug routes inflate the bill — every series you keep is a SKU.
Teams drop labels, shorten retention, and exclude services to control Datadog spend — gaps follow.
Drop at ingest, keep the signals you need — one project, worksheet-backed savings.
On Datadog a cardinality spike is an invoice surprise. On Elastic, Streams Drop caps what you store first.
On the Datadog worksheet, Datadog lands near ~$12.6K/mo while Elastic Serverless is ~$1.3K (89% cheaper) and ECH ~$0.6K. Metrics alone: Datadog ~$3.8K vs Serverless ~$0.9K (~4×). In the lab: Drop /debug/pprof, open Horizon — Your savings vs Datadog, then open the calc.
Full-stack list-rate (metrics + APM + logs + security) · Elastic Serverless is SaaS ($0 platform ops) · Metrics line: Datadog ~$3.8K vs Serverless ~$0.9K · tco-calc.o11ybot.app/?scenario=datadog · confirm with measured usage
Datadog custom-metrics pricing bills the spike first. Filtering later doesn’t undo the invoice.
Cut /debug/pprof and cardinality before ingest — noise never stored, bill stays predictable.
Same Streams pipeline behind both boards — no APM/traces required in this lab.
You already saw Datadog at ~$12.6K/mo vs Elastic Serverless ~$1.3K on the TCO worksheet (metrics alone ~$3.8K vs ~$0.9K). Now open Horizon — Your savings vs Datadog, Sample: Service Basics, and Horizon — Dashboard sprawl (hundreds of Datadog boards; custom metrics billed even if unused).
Quantify Drop + lifecycle — what Datadog would still bill as custom metrics.
Datadog export → live request/error metrics on Horizon without a rewrite project.
metrics-generic.otel-default — one stream, overrun control from Challenge 2.
Datadog dashboard sprawl: Elastic shows which boards query which indexes, and alerts if a schema change would blank a widget.
Bits investigates Datadog’s billed plane. Agent Builder investigates your Serverless project after Drop.
Open Agent Builder, ask about Horizon metrics on metrics-generic.otel-default, and leave with one clear contrast: AI that runs on metrics you chose to keep — not on the invoice you’re trying to leave.
Bits AI / Bits Agent Builder are strong — and still tied to metrics already stored on Datadog.
Same project as Streams Drop + metrics dashboards — Observability agent + skills, live on Serverless today.
“We Drop noise first, then Agent Builder investigates the cleaner, cheaper metrics in one project.”