Gigapipe is a self-hosted observability backend that stores logs, metrics, traces, and profiles in ClickHouse and serves them through drop-in Loki, Prometheus, Tempo, and Pyroscope APIs. It speaks LogQL, PromQL, and TraceQL, plugs into Grafana with no plugin, and its AGPLv3 core runs free with unlimited usage on operator hardware.[1][2]
Quick verdict
Self-hosted telemetry backend under Grafana Gigapipe gives Grafana shops a self-hosted ClickHouse-backed store behind familiar query APIs, trading SaaS convenience and built-in dashboards for data control and flat costs.[1][2][3]
Key features
Drop-in Loki, Prometheus, Tempo, Pyroscope APIs Serve LogQL, PromQL, and TraceQL behind the same endpoints those backends expose, so existing queries and alerts carry over.[1][2]
OpenTelemetry and Datadog Agent ingestion Collectors and the Datadog Agent point at Gigapipe directly, and a vendor comparison page documents side-by-side evaluation against Datadog.[1][2]
Influx and Elastic compatible senders Legacy Influx and Elastic shippers keep writing without a migration project on the sender fleet.[1][2]
ClickHouse plus S3-compatible storage All four signals land in operator-owned ClickHouse tables with object storage for scale.[1][2]
Grafana with no plugin Native Grafana datasources query Gigapipe directly; the project states it builds no dashboards itself.[1][2]
View explorer and CLI A built-in View explorer and CLI cover quick inspection without standing up a visualization layer first.[1]
Docker, Compose, Kubernetes, binaries Ship as a container image, a Compose full-stack example, a Kubernetes deployment, or amd64 and arm64 static binaries.[1][2]
Environment-variable operations Set basic-auth login, CORS origin, and retention TTLs through GIGAPIPE_ prefixed variables, with a documented Docker quickstart on port 3100.[1]
Who will love it
Grafana shops consolidating backends Grafana users wanting one self-hosted backend for logs, metrics, traces, and profiles can replace a per-signal backend sprawl with a single store.[1][2]
Teams escaping usage-metered billing Teams leaving per-signal SaaS billing can move to free self-hosted usage or flat-rate managed plans.[1]
On-prem data control On-prem environments can keep telemetry in owned ClickHouse and S3-compatible storage under their own control.[1][2]
Who should avoid it
External ClickHouse required Every deployment needs an external ClickHouse server, so the backend is only as solid as that dependency and teams without ClickHouse operations capacity should look elsewhere.[1][2]
No documented SSO No single-sign-on support appeared in the configuration reference, docs, or pricing pages checked; buyers requiring SSO should hold off.[1][2]
No vendor dashboards Gigapipe sits one layer below Grafana and builds no dashboards, so teams wanting an all-in-one surface from the same vendor should look elsewhere.[1]
How it fits your stack
Grafana-native query layer Grafana and Perses query Gigapipe through native Loki, Prometheus, Tempo, and Pyroscope datasource dialects with no plugin.[1][2]
Datadog Agent side-by-side path The Datadog comparison page documents native Datadog Agent ingestion for side-by-side evaluation during migration.[1]
Go service on operator hardware A Go service using ClickHouse client drivers and OpenTelemetry collector libraries runs on operator hardware via container or binary.[1][2]
Questions worth asking
What is the best Gigapipe alternative for cost-conscious teams?
Self-hosting Gigapipe itself is the lowest license-cost path, since the AGPLv3 core runs with unlimited usage on your hardware. If you need managed hosting without operating ClickHouse, compare its flat-rate Pro plan against Grafana Cloud and Datadog usage billing for your actual cardinality and retention.[1][2]
Is Gigapipe really free?
Does Gigapipe replace Grafana?
Sources and editorial review
Source research. Reviewed on October 10, 2026. Written by Sanjib Sah. AI assisted; reviewed by a human editor.
We have not run or tested this product.
- metrico/gigapipe GitHub repository
official · Observed October 5, 2026
Canonical repository metrico/gigapipe shows 1.7k stars, 103 forks, 2,308 commits, 20 issues, and an AGPL-3.0 license label; similarly named repos are forks of it.
- Gigapipe README
official · Observed October 5, 2026
README presents Gigapipe (formerly qryn) as a lightweight polyglot observability stack for logs, metrics, traces, and profiling with drop-in Loki, Prometheus, Tempo, and Pyroscope APIs.
- Gigapipe LICENSE
official · Observed October 5, 2026
Repository LICENSE file contains the GNU Affero General Public License version 3 terms.
- Gigapipe configuration reference
official · Observed October 5, 2026
Configuration reference documents Docker quickstart on port 3100 with a required external ClickHouse server and GIGAPIPE_ prefixed variables.
- Gigapipe go.mod
official · Observed October 5, 2026
go.mod declares a Go module using ClickHouse client drivers, Prometheus and OpenTelemetry collector libraries, and gRPC support.
- Gigapipe homepage
official · Observed October 5, 2026
Official homepage positions Gigapipe as a self-hosted replacement for Grafana Cloud, Datadog, or Elastic telemetry backends while keeping the Grafana UI.
- Gigapipe pricing
official · Observed October 5, 2026
Pricing page lists a free unlimited-usage AGPLv3 OSS tier, managed Pro from 349 EUR per month with 14-day trial, and Enterprise with HA, custom retention, and SLAs.
- Gigapipe Datadog alternative
official · Observed October 5, 2026
Datadog comparison page calls Gigapipe the open-source, OpenTelemetry-native alternative to Datadog with native Datadog Agent ingestion and a side-by-side table.
- Gigapipe Grafana Cloud alternative
official · Observed October 5, 2026
Grafana Cloud comparison page states Gigapipe does not build dashboards and works with Grafana and Perses one layer below as the telemetry store.
- Gigapipe OSS docs
official · Observed October 5, 2026
OSS docs give Docker, Docker Compose, Kubernetes, and static binary installs with ClickHouse as the required store and agent compatibility notes.