v0.7.2 — single-binary release

Unified observability, alerting,
and SIEM — self-hosted.

Replace Datadog, Splunk, and PagerDuty with one open-source stack you run on your own VPS. Metrics, logs, traces, alerts, and security events — under one binary, one schema, one dashboard.

Deploy in 5 minutes → View on GitHub curl -fsSL oleus.io/install.sh | sh
Trusted by teams shipping at scale 1,247 deployments · 38B events / day
/ 01 Capabilities

Six pillars. One binary. One schema.

Everything you need to know what your infrastructure is doing, and what's attacking it. No vendor lock-in. No per-host pricing. No data leaving your network.

M Metrics / 01

Prom-compatible at 1M samples/sec on a $48 box.

PromQL+MetricsQL. Long-term storage. Zero cardinality fees. Backfill from existing Prometheus instances on day one.

L Logs / 02

Structured logs at columnar speed.

SQL-native. JSON and OTLP ingest. Sub-second queries over a year of data. Up to 90% cheaper than Splunk-style stores.

SELECT count() FROM logs
WHERE service = 'api'
AND ts > now() - 1d;
→ 14,028,191 rows · 0.42s
T Traces / 03

Distributed traces, end-to-end correlated.

OTel-native. Cross-service spans linked to logs and metrics by trace_id. Tail-sampling included.

api · 248ms
db.query · 162ms
cache.get · 8ms
payments-svc · 118ms
A Alerting / 04

PagerDuty-grade routing — free.

PromQL alert rules. On-call schedules, escalations, deduplication, silencing. Notifies Slack, email, SMS, webhook.

High p99 on api-gateway
P1 · escalated → on-call rota
S SIEM / 05

Security events in the same pane.

File integrity, host IDS, log analysis, vulnerability detection. CIS benchmarks, PCI/HIPAA mappings out of the box.

P1SSH brute-force · 37.x.x.x
P2file modified · /etc/sudoers
P3CVE-2024-3094 detected
C CLI / 06

grep for your infrastructure.

Tail logs, query metrics, trigger silences, push dashboards — without ever opening a browser. Pipes into anything.

$ oleus tail --service=api | grep 5xx
$ oleus q 'rate(http[5m])' --since=1h
$ oleus alerts silence db.lag 30m
Full features breakdown →
/ 02 Deploy

From curl to dashboard in three steps.

A single static binary. No Kubernetes required. Runs on any Linux box from a $5 Hetzner CX11 to a 64-core bare metal.

01

Install the binary

One-liner. Verifies signature. ~24MB. No package manager required.

02

Provision the stack

oleus init spins up the full stack — metrics, logs, traces, alerting, and security — pre-wired with shared auth and a unified query layer.

03

Point your agents

OTLP, Prom remote-write, syslog, host agents — drop in. Existing Grafana dashboards keep working.

/ 03 Architecture

One stack. Your infrastructure.

No data leaves your VPC. Every layer is independently scalable, swappable, and inspectable.

ingest
OTLP
Prom remote-write
syslog · JSON
host agents
store
Metrics1M samples / s
Logscolumnar · SQL
TracesOTel-native
SecurityHIDS · FIM
act
alert engine
on-call routing
dashboards
oleus CLI
out
Slack · Email · SMS · PagerDuty webhook
S3 · Backblaze · MinIO (cold storage)
/ 04 Pricing

Self-hosted is free. Forever.

Oleus is Apache-2.0. Run it on your own boxes; pay only for the boxes. We make money offering managed hosting for teams who want to skip the ops.

Self-hostedfree · OSS
$0 / mo
  • All features. No host limits. No data caps.
  • Apache-2.0 license. Fork and modify freely.
  • Community support via GitHub & Discord.
  • Compatible with existing Grafana dashboards.
  • Zero data leaves your infrastructure.
Deploy in 5 minutes →
Managedcoming Q3
tbd / mo
  • Same product, hosted by us. SOC 2.
  • EU and US regions. Single-tenant available.
  • SLA-backed support. 24/7 on-call.
  • Migrate in or out at any time. Apache-2.0.
  • Volume discounts above 1B events/mo.
Try the cost calculator →
teams running oleus 1,247
avg. cost saved / mo $4.2k
events ingested / day 38B
github stars 14.2k
/ 05 ship it

Stop paying per-host.
Start running your stack.

One curl, one binary, one stack. Replace three vendors before lunch.

Deploy in 5 minutes → Browse all features