The News
ClickHouse has announced the public beta of ClickHouse Observability, a turnkey, serverless observability solution built for site reliability engineers and their AI agents, available on ClickHouse Cloud. The service ingests OpenTelemetry data through a single managed endpoint and delivers full-fidelity logs, metrics, and traces without sampling or roll-ups, paired with fast concurrent queries designed to support both human and agentic incident investigation. The launch is part of a broader platform expansion that also includes the general availability of Managed Postgres and the introduction of ClickHouse Security, positioning ClickHouse as a multi-workload data platform spanning transactional, analytical, observability, and security use cases.
Analyst Take
The full-fidelity bet
Most observability platforms solve the cost problem by discarding data. Sampling, roll-ups, and tiered retention are the industry’s default answers to the economics of telemetry at scale. ClickHouse Observability takes the opposite position: retain everything, make it queryable at sub-second latency, and let the cost model follow from the underlying engine’s columnar efficiency rather than from data reduction. That’s a meaningful architectural bet, and the Clever case study makes the argument concretely. A customer that previously indexed 10% of its logs and retained searchable data for three days now indexes 100% and retains 60 days’ worth, at no additional cost. That’s not a marginal improvement; it changes what investigations are possible.
For ITDMs evaluating observability spend, the pitch is straightforward: stop paying for data you deleted. The total cost of ownership argument is credible when the database engine doing the retention is ClickHouse’s columnar store, which was purpose-built for high-throughput analytical workloads. The risk is that “affordable retention” is a table-stakes claim in 2025, and buyers will rightly ask for benchmarks against Grafana Cloud, Datadog, and Honeycomb before accepting it at face value.
Agentic operations as the real design target
Read past the SRE headline and the more interesting signal is architectural. ClickHouse Observability exposes production telemetry to AI agents via the Model Context Protocol and SQL, effectively treating the observability layer as an MCP server that any compatible agent or model can query. That’s a forward-looking design choice. As agentic workflows move from experimentation into production operations, the observability platform needs to be queryable by machines, not just navigable by humans. Purpose-built investigation tools that bring code, runbooks, and operational context alongside telemetry are designed for exactly this pattern: an agent that can pull structured evidence, reason over it, and surface findings for a human to review.
Developers will recognize the practical implication immediately. If your incident response workflow involves an agent that can call MCP-compatible tools, ClickHouse Observability becomes a potential data source for that agent without custom integration work. The SQL interface lowers the barrier further. Whether the MCP ecosystem matures fast enough to make this a broad competitive advantage or a niche feature depends on adoption curves that remain genuinely uncertain, but the design intent is clear and it’s differentiated.
Where the platform story gets interesting
The simultaneous launch of Managed Postgres and ClickHouse Security alongside ClickHouse Observability signals something beyond a product feature release. ClickHouse is explicitly building toward a platform where transactional, analytical, observability, and security workloads coexist on a shared data infrastructure. That ambition puts it on a collision course with vendors like Datadog, which has been expanding from monitoring into security, and with cloud-native data platforms that are moving in the opposite direction, from warehousing toward operational use cases.
For engineering organizations, the consolidation argument has appeal. Managing fewer data platforms reduces operational complexity. According to ECI Research’s Google GovTech Survey, 21.8% of respondents cited “Too many disconnected security tools causing alert fatigue” as the biggest barrier to achieving a unified view of application security posture across multi-cloud or hybrid environments, and a separate 18.3% pointed to siloed development, operations, and security teams as the root cause. A platform that collapses observability and security telemetry onto a single queryable store directly addresses both of those pain points. The question is whether ClickHouse can execute across all four workload categories without the depth trade-offs that have historically plagued multi-workload platforms.
Developer velocity is also in the frame here. ECI Research’s Google GovTech Survey found that 47.2% of respondents selected “Developer velocity and ease of integration” as the factor carrying the greatest weight in their final technical selection process, assuming baseline security and compliance requirements are met. A serverless, OpenTelemetry-native service with a single ingest endpoint scores well on that dimension: there’s no infrastructure to provision, no pipeline to build, and the open standard means existing instrumentation carries over. That ease-of-integration story is ClickHouse’s strongest near-term sales argument.
Looking Ahead
ClickHouse Observability is in public beta, which means the next 6–9 months will be the proving ground. The company needs to demonstrate that full-fidelity retention at scale holds up operationally when customer data volumes spike, and that the AI-native investigation tooling delivers measurably faster mean time to resolution rather than just adding a notebook interface to a query engine. The MCP integration story, in particular, will depend on how quickly platform engineering teams at target customers actually deploy agentic operations workflows, a transition that is moving faster in cloud-native organizations than in regulated or government-adjacent environments.
The broader platform consolidation play is a longer arc. ClickHouse is building toward a position where it can credibly compete with Datadog on observability, with Snowflake on analytical workloads, and with purpose-built security data platforms, all simultaneously. That’s an ambitious surface area. The vendors most at risk in the near term are mid-tier observability platforms that lack ClickHouse’s cost efficiency story and haven’t yet built agentic interfaces into their products. ITDMs evaluating observability contracts in the next two quarters should put ClickHouse Observability on their shortlist, pressure-test the retention cost claims against their actual ingest volumes, and pay close attention to how the MCP and SQL interfaces fit their emerging agentic operations roadmap.
Stay Ahead of Application Development Trends
Get weekly analyst insights, research notes, event coverage, and AppDevANGLE updates delivered directly to your inbox.
Subscribe for Weekly Insights
Join technology leaders, practitioners, and GTM teams following the trends shaping modern software delivery.
Looking for deeper research access?
Explore ECI Research reports, survey insights, and market analysis through the ECI Research Portal.
