Unified observability, priced for the way engineers actually use it.
The wordmark of a company that would rather you never notice it. Last9 does its best work at 2 a.m., when a stream is peaking and nobody is thinking about the tool that's watching it.
There is a genre of software product that is boring in a way that is actually interesting, and Last9 is a good example. The company sells observability, which is the practice of watching your own software so you find out it's broken before your customers do. This sounds simple. It is not simple, and more to the point, it has become weirdly expensive.
Here is the problem in one word: cardinality. Every time your system emits a signal - a metric, a log line, a trace - it can carry labels. User ID. Region. Device. Version. Each new label multiplies the number of distinct series your monitoring system has to hold in its head. Modern applications, especially the consumer kind, emit telemetry with enormous dimensionality. Traditional time-series databases were not built for this, and when they hit it, they get slow, or they fall over, or they get so expensive that teams start deleting the very data they'd need during an outage.
Last9's founding observation was that most observability tools solve this by making you pay - either in dollars or in blindness. You either buy more capacity, or you pre-aggregate the detail away and hope the answer you need wasn't in the part you threw out. Last9's bet is that you shouldn't have to choose. Its telemetry warehouse, called Levitate, is a managed, Prometheus-compatible store built specifically to swallow high cardinality and concurrent access without the usual drama. You keep the detail. You keep the ability to slice by any label. And, ideally, you keep your budget.
The second piece is a Control Plane that sits in front of ingestion. Think of it as a bouncer for your telemetry: before data ever lands in storage - and on your bill - you can filter it, reshape it, turn logs into metrics, and route it wherever it needs to go. Crucially, you can do this at runtime, without re-instrumenting your services or shipping new code. When a spike in volume or cardinality shows up, you write a drop rule and the noise stops. This is a small idea with a large financial consequence, which is the kind of idea investors tend to like.
All of this rides on open standards. Last9 is OpenTelemetry-native and speaks PromQL, which means teams can adopt it without ripping out their existing dashboards or committing to yet another proprietary agent. Underneath, it leans on ClickHouse for cheap, fast columnar storage - a pairing the two companies have written about jointly under the banner of lowering the total cost of owning your observability.
The clearest proof that this works is a number that sounds made up: twelve of the eighteen largest live-streaming events in the world have run on Last9's observability, peaking at roughly 59 million people watching the same cricket match at the same time. When 59 million people press play and nothing breaks, no one sends a thank-you note. That silence is the product.
A unified observability control plane that correlates logs, metrics, and traces into one system for monitoring and faster triage - proven at 100+ billion events a day.
A managed, Prometheus-compatible time-series telemetry warehouse built for high cardinality, concurrent access, long retention, and proactive alerting.
Pre-ingestion filtering, reshaping, logs-to-metrics conversion, and routing - in real time, no re-instrumentation, with event-based pricing.
Model Context Protocol integration that pulls production telemetry into IDEs and coding agents, so root-cause analysis starts with real system context.
Last9's founders had already built and sold a company. They chose reliability - one of software's least glamorous problems - and went again. Boring problems have deep moats.
A repeat founder with roughly two decades in software product leadership. Previously co-founded Oogway Consulting, which was acquired in 2018 before he started Last9.
Modak's long-time collaborator and Oogway co-founder. Leads the engineering that makes a telemetry warehouse survive tens of millions of concurrent users.
Engineering, SRE, and DevOps teams at high-scale consumer and B2B companies - the kind that measure traffic in millions of concurrent sessions.
Backed by Sequoia Capital India's Surge program, Peak XV Partners, and Better Capital. Disclosed rounds total roughly $13M, against a reported ~$26.1M in total funding.
Modak and Verma co-found Oogway Consulting in Pune, focused on software scalability and reliability.
Oogway is acquired, setting up the founders' next reliability-focused venture.
Launches and joins Sequoia Capital India's Surge program with $2M in seed funding.
Raises $11M to scale the observability platform and the Levitate warehouse.
Powers observability at peaks near 59M concurrent viewers during major live-stream events.
Expands into a full telemetry data platform with cost-control routing and AI/MCP features.
Deepens its ClickHouse partnership to push down observability total cost of ownership.
It's a unified observability platform that correlates logs, metrics, and traces so engineering teams can monitor systems and triage incidents faster - with a strong focus on handling high cardinality affordably.
Levitate is Last9's managed, Prometheus-compatible time-series telemetry warehouse, built to handle high cardinality and concurrent access with long retention and proactive alerting.
Founded in 2020 by Nishant Modak (CEO) and Piyush Verma (CTO), repeat founders whose earlier venture, Oogway Consulting, was acquired in 2018.
It's OpenTelemetry-native and Prometheus-compatible, and leads with cost control: a pre-ingestion Control Plane and event-based pricing let teams filter and drop noisy telemetry in real time, without re-instrumentation.
Roughly $13M in disclosed funding - a $2M seed via Sequoia's Surge in 2020 and an $11M Series A in 2022 - with investors including Peak XV Partners and Better Capital.