Overview
A typical engagement starts with an engineering organization whose Datadog spend has become hard to forecast. The question is not only whether another stack would cost less. It is what a move would involve, what the team would have to run, and whether cleaning up the current bill would change the answer.
Challenge
Datadog bills each product on its own unit, so a single total hides which meter is growing. Dashboards and monitors do not convert automatically, instrumentation is tied to the Datadog agent, and the team has no one who has run Loki, Tempo or a long-term metrics store. Comparing prices across per-host, per-gigabyte and per-series models without the team's own volumes proves nothing.
Environment
A cloud estate of VMs and Kubernetes workloads with Datadog infrastructure monitoring, APM and logs, several hundred dashboards and monitors of mixed quality, and application teams that own their own instrumentation.
Approach
AceMQ engineers attribute the current bill to products and meters from the customer's usage views, then map each Datadog capability in use to a Grafana stack component or to a gap. They draft two designs, a self-run stack and a managed one, list what each would need operated, and write a parallel-run and cutover plan built on OpenTelemetry instrumentation. The write-up includes a cost-governance option for staying on Datadog.
Solution
- Attribution of the Datadog bill to hosts, containers, custom metrics, logs, APM and other meters, from the customer's own usage data
- Inventory of dashboards, monitors, integrations and instrumentation, each marked rebuild, retire or no equivalent
- Mapping of metrics, logs and traces to Prometheus or Mimir, Loki and Tempo, with Grafana for dashboards and alerting
- Two operating options written down: the customer runs the stack, or AceMQ runs it at Tier 3 or Tier 4
- OpenTelemetry-first instrumentation plan, then a parallel run with alert coverage proven equal before cutover
- A stay-and-clean-up option that targets custom metric tags, log indexing and host counts
Outcome
The engagement delivers a written assessment: the bill attributed by product, a map from each Datadog capability to a replacement or a gap, two operating options with what each requires, a parallel-run and cutover plan, and a clean-up option for staying. The customer decides from the facts, and AceMQ can run the replacement afterwards if they choose to move.
Technologies
Datadog
Grafana
Prometheus
OpenTelemetry
Kubernetes
Terraform
Linux
Amazon S3
