Learn OTel's Sum, Count, Histogram, and Last Value aggregation types with Python examples, plus delta vs cumulative temporality and best practices.
About us
Last9 is the preferred single pane of observability for Logs, Traces, Metrics, built for modern engineering teams and high cardinality environments. — OpenTelemetry native, Cloud Native, Prometheus compatible, with support for 100+ integrations — Simple pricing model with a predictable “pay for what you use” plan, no per host or per user billing ——— Application developers, infrastructure engineers, DevOps & SREs, and business users send their Logs, Traces, and Metrics to Last9’s telemetry data platform to monitor their services and systems and run alerting with anomaly algorithms for changes in patterns and behavior. Users are provided a Control Plane, as a first-class developer experience, to manage their telemetry data, its lifecycle, and its associated costs. Pre-Ingestion Workflows to filter, drop, and enrich data even before it’s stored don’t require instrumentation or code changes, and also enable engineering teams to not sacrifice visibility for cost optimizations. With Last9 built for high cardinality in mind — a generous 20M active time series per metric per day is supported, and tools like Cardinality Explorer & Streaming Aggregations allow identifying and handling telemetry with large cardinality. With correlated telemetry, Last9’s native explorer for Logs, Traces, and Metrics empowers users to quickly debug issues and not change context by allowing to jump between telemetry data, teams have the flexibility to use any tool that they’re comfortable with — either the embedded Grafana dashboard that Last9 comes with, or use Last9 as a data source for their preferred visualization tool. Organizations can choose between a SaaS or a BYOC/on-prem deployment model. Last9 is available on both AWS and GCP marketplaces. ——— Customers include Replit, Disney+ Hotstar, Clevertap, Circle, Probo, and Quickwork. Backed by Sequoia India and Better Capital, along with a handful of prominent angels and advisors from Hashicorp, Sendbird, GoJek, Razaorpay, and more.
- Website
-
https://last9.io/
External link for Last9
- Industry
- Software Development
- Company size
- 11-50 employees
- Headquarters
- Sunnyvale
- Type
- Privately Held
- Founded
- 2020
- Specialties
- Observability, Logs Management & Analysis, Application Performance Monitoring, High Cardinality Monitoring, and OpenTelemetry
Products
![Click here to view Last9 — Unified observability. Triage faster.]()
Last9 — Unified observability. Triage faster.
IT Infrastructure Monitoring Tools
Last9 provides unified observability at scale, to triage faster. With a single pane for correlated telemetry, engineering teams can jump from Logs to Traces to Metrics to quickly debug issues and implement fixes, powered by a telemetry warehouse with a first-class Control Plane that enables users to manage their telemetry data and its lifecycle.
Locations
-
Primary
Get directions
Sunnyvale , US
Employees at Last9
Updates
-
Last9 reposted this
NVIDIA, AMD, and Intel Gaudi in one fleet, on one dashboard. We made that a hard design rule, not an afterthought. Two decisions that matter more than any single metric we collect: 1) One metric namespace across vendors. NVIDIA's NVML, AMD's amd-smi, and Gaudi's hl-smi all normalize to the *same* OTel names and units. Add AMD capacity next quarter and your dashboards and alerts don't break. 2) No vendor backend in the agent. On purpose. It emits standard OTLP: Datadog, Grafana, Honeycomb, New Relic, @last9, whatever you run. The agent's job is to produce clean, attributed telemetry, not to trap it. → Locked: telemetry that only works inside one vendor's walls. → Neutral: telemetry that follows your hardware *and* your backend choices. The GPU hardware mix is going to keep changing. Your observability shouldn't have to. If you're building a platform: how are you handling a multi-vendor GPU future today? #OpenTelemetry #GPU #ArtificialIntelligence #AIInfrastructure #Technology
-
-
#newinlast9 A round of improvements across our AI Assistant, MCP server, and log ingestion. → AI Assistant — a refreshed chat experience: denser layout, syntax-highlighted tool inputs and outputs, and a compact tool-call timeline. → MCP — the AWS DevOps Agent can now connect to your hosted MCP server (registered as a standard OAuth PKCE client); And when you ask about structured fields like HTTP status codes, the assistant now filters on the attribute itself instead of doing a plain-text search — so you get the right results. → Integrations — logs ingested from Amazon S3 now always arrive with a service name, with built-in parsing for AWS WAF and ALB logs. Plus readable durations (ms/s) in queries, MS Teams Power Automate notifications, and fixes for Kubernetes pod-state lag, dashboard table sorting, and RUM error views.
-
-
Last9 reposted this
Standard OpenTelemetry Java agent works great for most Java Apps except Vert.x We open sourced our work after testing in production to support Zero-code OpenTelemetry auto-instrumentation for Vert.x https://lnkd.in/gjksVfnt
-
Setting up OpenTelemetry for a Go service is usually a few hundred lines - exporters, tracer providers, propagators, the works. Or: agent.Start() Then swap grpc.NewServer() for grpcagent.NewServer(), and every RPC is traced. Same drop-ins for fasthttp, Gin, and SQL. Standard OTel under the hood. Open source - link in comments
-
-
Your observability stack just got a lot smarter. Catch Last9 at the Asia DevOps Conference — Kuala Lumpur, June 10. Showing agent monitoring, coding-agent monitoring, MCP, and our in-app AI assistant, alongside our unified observability platform for logs, metrics, and traces. Pullman Kuala Lumpur City Centre • June 10, 2026 We'd love to meet you at the booth. 🔗 https://lnkd.in/gYNJCCmg Manuel Oothria Raj Apurva Chaudhary Sahil Khan Nishant Modak Shekhar Patil Kuldeep Dhankar Pournami RS Mukta Aphale Prathamesh Sonpatki
-