Here’s something most observability bills hide: A large share of what you pay to ingest, nobody ever looks at. Debug logs left on from a deploy three months ago. Health checks firing every few seconds. The same error, repeated thousands of times, each copy billed like it’s news.
On November 12, Gartner® published its Magic Quadrant™ for Observability Pipelines, and Apica was positioned in the Challengers Quadrant. We believe that positioning reflects an idea we’ve built around from day one: enterprises don’t need to ingest everything to understand everything.
The press release covers the news. Here I want to explain why we think the pipeline, not the backend, is where observability costs get decided.
Ingest everything is a bill, not a strategy
For years, observability ran on a simple deal: Send all of it to the backend and work out what matters later. That deal worked when data grew at a predictable pace. It doesn’t work when AI agents are in the mix.
Agents don’t make one call and stop. They reason, retry, call tools, call other agents, and log every step. In research we commissioned from Omdia, enterprises projected telemetry growth of 9.5x. The average organization already spends $3.17 million a year on observability. And 59% said they had delayed or terminated agentic AI deployments because of monitoring costs.
That last number is the one that worries me. When the cost of watching your AI is high enough to stop you from shipping it, the problem isn’t your AI. It’s the ingest-everything model.
Route before you ingest
The cheapest byte is the one you never send to an expensive backend. That’s the job of Apica Flow, our telemetry pipeline. Flow sits between your sources and your destinations, and it decides what happens to every piece of telemetry before you pay to store or index it:
- Filter the noise. Drop the debug logs, duplicate events, and health checks no one queries.
- Trim and mask. Strip fields you don’t need and mask sensitive data at the point of ingestion, before it lands anywhere.
- Enrich what’s left. Add the context that makes the remaining data more useful, so you need less of it.
- Route by value. Send high-value data to Datadog, Splunk, or your SIEM. Send the rest to lower-cost storage, where it stays available to replay if an investigation needs it.
That last point matters. Reducing data is easy if you’re willing to lose it. Reducing it while keeping the option to bring it back is what lets teams cut aggressively without fear.
The result: Apica customers can lower the total cost of ownership of observability by up to 40% compared with legacy observability platforms. Your teams keep the tools they know. They just stop paying those tools to store data nobody needs.
A pipeline you can trust in the critical path
Putting anything between your systems and your observability tools is a real decision. If the pipeline fails, you lose visibility exactly when you need it most. So, we built Flow around a simple rule: Never block, never drop.
When a destination slows down or goes offline, Flow buffers telemetry persistently and replays it once the destination recovers. Your sources keep sending, and accepted data stays intact through the interruption.
A few other choices follow from the same thinking:
- OpenTelemetry-native. Flow works with open standards, so you aren’t locked into a proprietary agent or format.
- Fleet management alongside Flow. Manage collection agents across distributed environments from one place and push policy changes without touching each host.
- Your tools of choice. Flow sends data to the observability and security tools you already run. No rip and replace required.
Humans and agents, on your terms
AI belongs in the pipeline. But I don’t think any vendor should decide which AI you use, or where your telemetry goes to reach it.
With Apica, you bring the AI tools of your choice, including models you host yourself. Your telemetry doesn’t have to pass through yet another vendor’s AI service. And no AI-initiated action can delete data or materially change configuration without a human approving it first.
As a Swedish-American company, we take data sovereignty seriously. You own your telemetry, and you own your AI choices. That holds whether the next request comes from an engineer or an agent.
Where the cost gets decided
Observability costs aren’t decided at the backend. By the time data gets there, you’ve already paid for it. They’re decided in the pipeline, at the moment you choose what to keep, where to send it, and at what fidelity.
We believe our positioning in the Gartner Magic Quadrant for Observability Pipelines reflects that focus. And we’re not done. AI agents will keep multiplying the data they generate, and we’ll keep building the pipeline that keeps that data under control.
If your observability bill is growing faster than your visibility, let’s talk. Read the Gartner Magic Quadrant for Observability Pipelines
Gartner, Magic Quadrant for Observability Pipelines, Pankaj Prasad, Andre Bridges, Neil Young, Andrea Curry, 12 November 2026.
Gartner and Magic Quadrant are trademarks of Gartner, Inc. and/or its affiliates.
Gartner does not endorse any company, vendor, product or service depicted in its publications, and does not advise technology users to select only those vendors with the highest ratings or other designation. Gartner publications consist of the opinions of Gartner’s business and technology insights organization and should not be construed as statements of fact. Gartner disclaims all warranties, expressed or implied, with respect to this publication, including any warranties of merchantability or fitness for a particular purpose.
