From DataDog to ClickStack: Notes From a Real Observability Migration

Good observability does two things at once: it gives engineers a clear, fast path from a symptom to its root cause, and it scales with your data without forcing hard tradeoffs about what you can afford to watch. When telemetry volumes grow, that second part gets harder, and the architecture underneath your monitoring stack starts to matter as much as the features on top of it.

On a recent engagement, one of our clients - Rosen Publishing - was running on DataDog, a capable and widely used platform. As their data volumes and cardinality grew, they wanted an approach that scaled more naturally with that growth. So we migrated their observability stack to Managed ClickStack, the observability platform from ClickHouse, our partner in this space. The migration was straightforward, and the result was simpler to operate than what it replaced. Here is what we found.

What ClickStack Is Built For

ClickStack is built on ClickHouse, a columnar database engineered for fast analytics on massive, high-cardinality datasets. That foundation is the whole point. High-cardinality data, the kind generated by Kubernetes clusters and per-user attributes, is exactly what ClickHouse handles efficiently, so visibility scales with your systems instead of working against you.

Why ClickStack

It bundles an OpenTelemetry collector for ingestion with the HyperDX UI for visualization, and it stores logs, traces, metrics, and session data together as wide events in a single store, fully queryable and correlatable at the database level.

We chose the managed offering rather than running the stack ourselves. With Managed ClickStack, ClickHouse and the HyperDX UI are hosted and operated in ClickHouse Cloud, so the only piece our client runs is the OpenTelemetry collector, which deploys as a Docker container. That kept the operational surface small: standard OpenTelemetry instrumentation in, a single container to manage, and everything else handled by the platform.

Columnar storage and strong compression also meant a meaningful reduction in storage footprint, which keeps the economics predictable as data grows.

What Actually Changed

A few differences stood out once the migration was complete.

Everything we tracked before, in a simpler interface. ClickStack surfaced all of the same telemetry Rosen Publishing relied on, logs, metrics, and traces, with a more streamlined experience. Nothing essential was lost in the move, and day-to-day debugging got easier.

Predictable economics at scale. Because the data sits in a store built to hold high-cardinality telemetry efficiently, cost tracks data volume in a way that stays predictable as systems grow, which keeps visibility and budget aligned.

Open querying. HyperDX supports Lucene-style search for fast exploration, and because the data sits in ClickHouse, engineers can drop into SQL for any analytical question. There is no proprietary query language to learn, and no question that is off-limits.

OpenTelemetry-native ingestion. Instrumentation is standards-based, so the telemetry stays portable and there is no lock-in to a single vendor's agent.

End-to-end tracing got simpler. With everything in one store, there was no stitching telemetry together across separate tools using timestamps or correlation IDs.

The Migration Itself

The move was straightforward. Because we ran the managed offering, there was no database or UI to stand up and operate ourselves, just the OpenTelemetry collector in a Docker container pointed at ClickStack. The decisions that still mattered were the usual ones: retention policies, alert structure, dashboard parity, and team onboarding. Getting those right is what separates a clean cutover from a noisy one, and it is where most of the real engineering value lives.

How Lineate Can Help

We have deep experience building data-centric systems on ClickHouse, and our partnership with ClickHouse keeps us close to where ClickStack is heading. If you are looking for observability that scales naturally with your data, we can help you evaluate ClickStack against your current setup and plan a migration that preserves what already works for your team.