2,431,902 events/sec flowing now

Real-time metrics, the moment they happen.

DataPulseStream ingests millions of events per second and makes them queryable in milliseconds — so your dashboards, alerts, and traces are never a step behind production.

$ npm i @datapulse/stream
ingest · us-east-1 LIVE
events / sec 2.43M  ▲ 4.1%
Ingest rate
2.43M/s
p99 query
41ms
Active streams
8,204
event stream
Powering observability for teams shipping at scale Northwind Halyard Quanta Labs Meridian Voxel CI
The platform

Everything you need to watch production in real time.

Built by engineers who got tired of dashboards that lag a minute behind the incident. No batching delay, no sampling surprises — just the truth, as it streams in.

Ingest at any scale

Drop-in agents, native OpenTelemetry, and SDKs for every runtime. From one container to ten thousand pods — no config rewrite.

OTLP · Prometheus · StatsD · syslog

Query in milliseconds

A columnar engine purpose-built for time series. Slice billions of points across any dimension without watching a spinner.

p99 < 50ms · PromQL + SQL

Alerts that fire on the stream

Rules evaluate on the live pipeline, not a cron. Page the right person within a second of the spike — before customers feel it.

PagerDuty · Slack · webhooks

Engineer-first by design

Everything is an API. Version your dashboards, script your alerts, drop it all in CI. No clicking through onboarding wizards.

Terraform · CLI · GitOps-ready
How it works

Three steps from emit to insight.

No collectors to babysit, no warehouse to provision. Point your services at one endpoint and start querying within minutes.

01

Emit

Your services push events over the SDK or any OTLP endpoint. One line of code, fire-and-forget.

pulse.emit("checkout.completed", …)
02

Stream

We ingest, deduplicate, and index on the fly. No batching window, no warehouse load job standing between you and the data.

ingest → index → fan-out  <100ms
03

Observe

Query, chart, and alert on data that's milliseconds old. Build dashboards or wire it into your own tools.

SELECT p99(latency) FROM events
Developers

Instrument in the language you already ship.

Typed SDKs, sensible defaults, and an HTTP API when you want raw control. Send your first event before your coffee's cold.

import { Pulse } from "@datapulse/stream";

const pulse = new Pulse({ apiKey: process.env.DPS_KEY });

pulse.emit("checkout.completed", {
  amount_cents: 4200,
  region: "eu-west-1",
  latency_ms: 38,
});
from datapulse import Pulse

pulse = Pulse(api_key=os.environ["DPS_KEY"])

pulse.emit("checkout.completed", {
    "amount_cents": 4200,
    "region": "eu-west-1",
    "latency_ms": 38,
})
pulse := datapulse.New(os.Getenv("DPS_KEY"))

pulse.Emit("checkout.completed", datapulse.Fields{
    "amount_cents": 4200,
    "region":       "eu-west-1",
    "latency_ms":   38,
})
curl -X POST https://api.datapulsestream.com/v1/ingest \
  -H "Authorization: Bearer $DPS_KEY" \
  -d '{"event":"checkout.completed",
      "amount_cents":4200,"region":"eu-west-1"}'

Zero-config batching

The SDK buffers, compresses, and retries for you. Backpressure handled — your hot path stays fast.

Typed end to end

Schemas inferred from your events and enforced at query time. Catch a renamed field in CI, not at 3am.

OpenTelemetry native

Already on OTel? Point your exporter at our endpoint and keep your existing instrumentation.

2.4M/s
Sustained ingest per region, with headroom to burst.
<50ms
p99 query latency across billions of points.
99.99%
Uptime SLA, multi-region by default.
14days
Full-resolution history on the free tier.
Get started

Start streaming in under five minutes.

Create a key, send your first event, watch it land on a live chart. No credit card, no sales call — just you and your data.

Free tier · 1M events/day · upgrade when you outgrow it