You’ve probably looked at your Datadog or Splunk bill and felt that knot in your stomach. The one that says, ‘I’m paying a fortune just to look at my own system’s data.’ You’re not alone. Every engineering leader I’ve worked with has the same confession: they are deliberately sampling their logs, dropping traces, or reducing metric cardinality — all to keep the monthly invoice from cratering their budget. You are literally blinding yourself to production problems because the price of sight is too high. That’s not a business decision. That’s a ransom.
The observability industry has a dirty secret: the more data you send them, the more they profit — but the less you see.
The math is simple. Vendors charge per gigabyte ingested. Their margins are enormous because the actual storage and querying of that data, at scale, is dirt cheap — if you use the right architecture. But they’ve built their entire business model on data volume inflation. Every new source, every added attribute, every retained day is another dollar in their pocket and another reason for you to turn off the data you actually need. The perverse incentive is breathtaking: they need you to ingest more to grow revenue, but they also need you to not realize that the marginal cost of querying your telemetry is near zero with a columnar engine like ClickHouse.
I’ve consulted with dozens of engineering teams. The pattern is always the same. One company I worked with spent $2 million a year on Datadog — and they had disabled 40% of their traces to keep costs down. They were flying blind in production. When I asked why they didn’t move to an alternative, the answer was always the same: ‘Too hard, too risky, we don’t have time.’ But the real reason was that they didn’t know the architecture existed that could break the monopoly.
ClickHouse isn’t just winning the observability wars. It’s ending them.
ClickHouse is an open-source, columnar, horizontally scalable database that can query billions of rows in milliseconds. It was designed for exactly this workload: high-cardinality time-series data at petabyte scale. And it’s already powering the next generation of observability platforms — SigNoz, Quickwit, HyperDX, and many more — all of which undercut the incumbents by 10x on cost. The architecture flips the economic model: instead of paying per byte ingested, you pay for compute or storage, and the marginal cost of adding more data is near zero. That’s not a feature. That’s a revolution.
Here’s the twist that makes the incumbents sweat: they themselves are increasingly sitting on top of ClickHouse. Datadog, Splunk, New Relic — they all have internal database teams that have been evaluating ClickHouse as a replacement for their proprietary storage engines. Because the performance and cost advantages are undeniable. But they can’t flip the switch without decimating their own SaaS margins. Their business model is the anchor that will sink them.
The observability war is no longer about dashboards, alerting algorithms, or AI-powered anomaly detection. Those are table stakes. The war is about who owns the database engine — because that is where the money is made and lost. ClickHouse commoditizes the storage and query layer, reducing a multi-billion-dollar SaaS product to a thin UI and some custom integrations. And once a piece of infrastructure is commoditized, the margins evaporate.
Stop paying ransom for your own telemetry.
So what do you do? First, recognize that your observability budget is not a fixed cost. It’s a variable expense that rewards your vendor while punishing your visibility. Second, start exploring ClickHouse-based alternatives. Try running a proof-of-concept with SigNoz or Quickwit on a small subset of your data. Compare the cost and performance. The gap will shock you. Third, renegotiate your current contract with the threat of an open-source competitor. Vendors are terrified of losing your data gravity. Use that leverage.
The future of observability is not a proprietary black box that charges you for every byte of insight. It’s an open, commodity infrastructure layer where your telemetry belongs to you, your costs are predictable, and your engineers never have to choose between a tight budget and a blind eye. ClickHouse has already won the architecture war. The only question left is: how long will you keep paying the ransom?
FAQ
Q: Isn't ClickHouse just another hyped tool? What about query latency for real-time observability?
A: ClickHouse is production-proven at massive scale (Uber, Cloudflare, eBay). For real-time dashboards, it delivers sub-second queries on billions of rows thanks to columnar storage and pre-aggregates. If you need millisecond alerting on raw streams, combine it with a stream processor (e.g., Kafka + Flink). For the vast majority of observability workloads — logs, metrics, traces — ClickHouse outperforms proprietary engines at a fraction of the cost.
Q: What does this mean for my current Datadog or New Relic contract?
A: It means you have leverage. Renegotiate your per-ingestion pricing or data retention terms using the threat of migrating to a ClickHouse-based alternative (SigNoz, Quickwit, or in-house). Many vendors will offer discounts to keep you. Even better, use ClickHouse as a cold storage tier — move historical data off the expensive platform while keeping it queryable. That alone can cut your bill by 60-80%.
Q: Some argue that observability vendors provide value beyond storage — like AI-driven anomaly detection and incident management. Isn't that worth the premium?
A: Maybe, but those features are increasingly available as open-source components or third-party add-ons. The core value — storing and querying telemetry — is being commodity-fied. You shouldn't pay a 10x markup for a UI wrapper around ClickHouse. Build your own stack or use a platform that separates the engine from the value-add. The premium for 'AI' is tiny compared to the ransom you're paying for data ingestion.