Observability: Channeling Telemetry

Not ranked, but frequently cited: that is ClickHouse’s position in the latest Magic Quadrant for observability platforms.

Over the past year, the American company has been evolving in this market. Building on the HyperDX acquisition, it actually launched ClickStack, integrated with its database engine. Early 2026 brought another acquisition — Langfuse — which added application monitoring capabilities and AI agents tailored for workloads.

This product earned it an “Honorable Mention.” Gartner also grants one to Dash0, founded by the creators of Instana… whose platform is built on ClickHouse. The same distinction goes to groundcover, whose “bring your own cloud” architecture lets telemetry be stored on private ClickHouse instances.

A new criterion — optional — for telemetry filtering

From year to year, the list of functional criteria to qualify for Gartner’s Magic Quadrant for observability has evolved only modestly. The criteria that remain mandatory are:

• Ingestion, storage and analysis of telemetry
• Identification and analysis of the behavior of applications, services and infrastructure
• Telemetry enrichment, notably through dependency mapping and relationships between services

Gartner added the collection of telemetry from public cloud providers “such as AWS and Azure.” It also includes automated discovery and mapping of components and related infrastructure/network/application services.

This latter element was optional last year. It was more precisely part of a set of criteria that vendors had to cover at least four of. That requirement is no longer in effect: the criteria remain, but they are simply “optional.” Among them:

• Integration of ITSM, CMDB, automation, DevOps tools…
• Cost management
• Observability of AI workloads
• Automation of code and configuration changes
• Identification of vulnerabilities and blocking exploitation attempts

This year, Gartner added telemetry filtering/sampling. Behind this, a trend it has flagged for several editions of this Magic Quadrant: the difficulty of managing costs. Several leaders receive comments on this aspect, sometimes positive, sometimes negative.

Splunk is no longer a “leader” in observability

Vendors were evaluated on two aspects. One, “execution,” reflects their ability to meet demand (customer experience, pricing, quality of products/services…). The other, “vision,” reflects their strategies (product, innovation, marketing…).

The situation on the “execution” axis:

On the “vision” axis:

Seven of the eight “leaders” from last year remain in leadership positions (Chronosphere, Datadog, Dynatrace, Elastic, Grafana Labs, IBM and New Relic). Splunk is an exception, demoted to the “challengers” due to a retreat on the “vision” axis. Coralogix replaces it. It had been among the “visionaries” in 2025.

Classed as “niche players” last year, ITRS and Sumo Logic exit the Magic Quadrant. Oracle fares similarly, having been a “challenger.” HPE breaks into the “niche players”; Alibaba Cloud lands among the “challengers.”

Chronosphere acquired by Palo Alto Networks: an integration to watch

Chronosphere has belonged to Palo Alto Networks since January 2026, and the plan is to keep it as an autonomous offering. It stands out for its ability to handle high-volume telemetry while preserving cost control. Gartner also salutes its latest AI-agent innovations for problem identification. It also notes what the Palo Alto acquisition brings in terms of business opportunities and financial footing.

The impact of this acquisition on Chronosphere’s activity will be watched closely. In particular, its support model — which has been a distinguishing feature — and whether joining Palo Alto will broaden its direct presence beyond a largely North American footprint. Gartner also notes that Chronosphere may not be the best fit for legacy APM needs (Java, .NET).

At Coralogix, the customer is responsible for the S3 infrastructure

Gartner appreciates Coralogix’s real-time telemetry analysis and the potential cost savings that follow (the ability to leverage lower-cost storage classes). As with Chronosphere, it highlights innovations in agentic AI (Olly agent, MCP server, Slack and GitHub integrations). It also points to granular cost management by telemetry type with priority levels.

Coralogix suffers from limited visibility and its product is still perceived as “log-first.” Its direct presence is more limited than other leaders in the APAC region and in Latin America. Attention also to the underlying S3 infrastructure, which the client must manage (usage and costs).

OpenTelemetry, not at the heart of Datadog’s value proposition

The breadth of Datadog’s portfolio (“more than 20 products within one interface”) resonates with Gartner. The Bits AI SRE component also signals maturity in deploying GenAI. And in marketing, with its own annual DASH conference as a focal point.

The granularity of pricing for ingestion, indexing, hosts and containers requires careful oversight. Moreover, while Datadog supports OpenTelemetry, its value proposition remains largely dependent on its proprietary agent. As for the “bring your own cloud” option, it covers logs, but not metrics or traces.

Dynatrace and its learning curve

Dynatrace’s compliance focus earns it a solid point. The real-time map of dependencies, supported by a deterministic framework that automates root-cause isolation, also earns praise. Gartner also values the AI layer that helps define objectives bounded by organizational guardrails.

Watch the onboarding: getting full value from Dynatrace can require a fairly steep learning curve. Gartner also recommends monitoring ongoing discussions with an activist investor (Starboard Value, which took a majority stake in April). And it regrets AI marketing that can confuse buyers.

Tuning work with Elastic

The triad of search, security, and observability remains compelling — especially economically — for those already in the Elastic ecosystem. Another strength is the level of OpenTelemetry integration, between the catalog of community collectors and supplier-specific distributions. Gartner also applauds advances in agentic AI (agent builders, skills library).

Elastic is often seen primarily as a search engine. Its query language requires substantial configuration to be functionally equivalent to competitors on features like topology mapping and automated remediation. Also beware of cost control, especially for self-managed deployments when dealing with large indexes.

With Grafana Labs, flexibility can breed complexity

Grafana Labs’ open-source heritage and its strong alignment with Kubernetes environments provide a “frictionless entry,” in Gartner’s words. The firm is also praised for its automated telemetry filtering and its contextualization via a knowledge graph.

Grafana offers flexibility to assemble observability stacks, but it demands expertise in Prometheus and OpenTelemetry. The fact that it uses multiple back-ends (Mimir for metrics, Loki for logs, etc.) can introduce architectural complexity. Gartner also notes the cyberattack the company recently suffered (GitHub intrusion following credential theft as part of the Mini Shai-Hulud campaign).

IBM has not yet unified Instana, Turbonomic and SevOne

Beyond the commercial opportunities that come with its installed base in other markets, IBM stands out for integrating Kubecost, lending a FinOps angle to Instana. Gartner also appreciates its “transparent approach” that applies only deterministic models to identify problems while reserving generative models for explanations and recommendations.

Under the Concert brand, IBM promises a more unified experience across Instana, Turbonomic and SevOne — which remain, for now, distinct products with separate installation and administration. Gartner also points to the risk of “brand dilution” for Instana within the group’s portfolio. And licensing that can be “often complex” when embedded in IBM’s enterprise agreements.

New Relic: with AI, a business model worth studying

New Relic earns Gartner’s credit for a solid AI agentic layer (used for automated root-cause analysis and log summarization). It is also praised for telemetry filtering and sampling capabilities. Its partnerships (Accenture and NTT Data are named) have extended its reach.

New Relic tends to lose ground to competitors in the AI space. It also lacks a flagship event, even though it has announced the return of its annual conference in 2026. It is important to carefully study the implications of the Compute Capacity Units credit system introduced for AI features.