Self-hosted observability platform built in Rust
Logs, traces, errors, and server metrics — all in one place.
Features • Quick Start • Sending Data • Frontend Errors • Metrics Agent • Configuration
Most observability tools are either expensive SaaS products or complex self-hosted stacks requiring Kafka, Elasticsearch, and a PhD in YAML configuration.
App Beholder is different: one binary, one database, zero complexity. Deploy it in minutes and start seeing what your apps are doing.
- No vendor lock-in — your data stays on your infrastructure
- No complex pipelines — just POST JSON or use OpenTelemetry
- No heavy dependencies — just PostgreSQL
- Real-time streaming — logs appear instantly via SSE
Real-time scrolling log view with level, host, and text filtering. Logs stream in via Server-Sent Events as they arrive. Click any log entry to see full attributes and stack traces. Jump from a log directly to its trace.
Waterfall visualization of distributed traces. See the full span tree with timing, parent-child relationships, and error highlighting. Drill into any span to see attributes and linked logs.
CPU, memory, swap, disk, network, and per-process metrics charted over time with multi-host support.
Track multiple applications across multiple servers. Projects are auto-created on first event — no setup required. Per-host stats with log and span counts.
| Layer | Technology |
|---|---|
| Backend | Rust, Axum, tokio-postgres |
| Frontend | HTMX, SSE, Tailwind CSS, DaisyUI |
| Charts | uPlot |
| Database | PostgreSQL (time-partitioned tables) |
| Templates | Askama (compile-time) |
- PostgreSQL 14+
- Linux (amd64 or arm64) — for prebuilt binaries
- Rust 1.75+ — only if building from source
createdb appbeholderEvery push to main publishes ready-to-run binaries. No Rust toolchain required.
# Pick your architecture: linux-amd64 or linux-arm64
ARCH=linux-amd64
# Download the latest release
gh release download --repo mi4uu/appbeholder --pattern "appbeholder-${ARCH}" --pattern "appbeholder-agent-${ARCH}" --dir .
# Make executable
chmod +x appbeholder-${ARCH} appbeholder-agent-${ARCH}
# Run the server
./appbeholder-${ARCH}Or download manually from the Releases page.
Each release includes four binaries:
| Binary | Description |
|---|---|
appbeholder-linux-amd64 |
Server (x86_64) |
appbeholder-linux-arm64 |
Server (aarch64/Graviton) |
appbeholder-agent-linux-amd64 |
Metrics agent (x86_64) |
appbeholder-agent-linux-arm64 |
Metrics agent (aarch64/Graviton) |
git clone https://github.com/mi4uu/appbeholder.git
cd appbeholder
cargo build --release -p appbeholder
cargo build --release -p appbeholder-agent
# Run the server
./target/release/appbeholdercurl -X POST http://localhost:8080/api/v1/logs \
-H "Content-Type: application/json" \
-H "X-Project-Slug: my-app" \
-d '{"level":"info","message":"Hello from App Beholder!"}'Open http://localhost:8080 — your project appears automatically and the log is streaming live.
All endpoints require the X-Project-Slug header to identify the project. The host is auto-detected from the request IP, or you can set it explicitly with X-Host.
{
"level": "error",
"message": "Failed to connect to database",
"timestamp": "2026-03-18T12:00:00Z",
"trace_id": "abc123",
"attributes": { "module": "db", "retry_count": 3 },
"stack_trace": "Error: connection refused\n at db.connect..."
}| Field | Type | Required | Description |
|---|---|---|---|
level |
string | yes | debug, info, warn, error, fatal |
message |
string | yes | Log message |
timestamp |
ISO 8601 | no | Defaults to server time |
source |
string | no | backend (default) or frontend |
trace_id |
string | no | For linking to traces |
span_id |
string | no | For linking to spans |
attributes |
object | no | Arbitrary key/value metadata |
stack_trace |
string | no | Full stack trace |
Convenience endpoint — creates a log entry with level=error and auto-generates an error fingerprint for grouping.
{
"message": "TypeError: Cannot read property 'id' of null",
"stack_trace": "TypeError: Cannot read property...\n at UserService.getUser...",
"source": "frontend",
"attributes": { "url": "/dashboard", "browser": "Chrome 120" }
}Batch endpoint for system metrics (used by the agent).
{
"metrics": [
{ "name": "cpu_usage_total", "value": 72.5, "unit": "percent" },
{ "name": "memory_used", "value": 3221225472, "unit": "bytes" },
{ "name": "process_cpu_usage", "value": 45.2, "unit": "percent",
"attributes": { "pid": "1234", "name": "postgres" } }
]
}Point your OpenTelemetry SDK exporter at App Beholder:
OTEL_EXPORTER_OTLP_ENDPOINT=http://your-server:8080
Supported endpoints:
POST /v1/traces— OTLP trace ingestionPOST /v1/logs— OTLP log ingestion
Add a single script tag to automatically capture frontend errors:
<script src="http://your-server:8080/static/beholder.js" data-project="my-app"></script>This automatically captures:
window.onerror— uncaught exceptionsunhandledrejection— unhandled promise rejections
Errors are sent to POST /api/v1/errors with source: "frontend" and grouped by fingerprint.
The appbeholder-agent binary collects system metrics and sends them via OTLP/HTTP.
# Using prebuilt binary
./appbeholder-agent-linux-amd64 agent.toml
# Or build from source
cargo run -p appbeholder-agent -- agent.tomlCreate an agent.toml:
endpoint = "http://localhost:8080"
service_name = "my-app"
hostname = "web-1"
interval_secs = 30Collected metrics:
- CPU — total usage, per-core usage, load average (1m/5m/15m)
- Memory — total, used, available
- Swap — total, used, free
- Disk — total, used, available, read/write throughput per mount
- Network — RX/TX bytes per interface
- Processes — top 10 by CPU and memory (PID, name, usage)
Metrics are sent every 10 seconds by default.
App Beholder uses a simple password gate. Create a .password file next to the binary:
echo "my-secret-password" > .password- If
.passwordexists: web UI requires login, API requiresX-Api-Passwordheader - If
.passworddoesn't exist: no authentication (open access)
Create a config.toml in the working directory:
[server]
host = "0.0.0.0"
port = 8080
[database]
url = "postgres://localhost/appbeholder"
[retention]
logs_days = 7 # How long to keep logs
traces_days = 30 # How long to keep traces
metrics_days = 90 # How long to keep metrics
errors_days = 30 # How long to keep error groupsAll values have sensible defaults — the config file is optional.
+-------------------+ +--------------------+
| Client Apps | | appbeholder-agent |
| (OTLP / REST) | | (system metrics) |
+---------+---------+ +---------+----------+
| |
v v
+-------------------------------------------------+
| App Beholder Server |
| +-----------+ +-----------------------------+ |
| | OTLP HTTP | | REST API | |
| | Receiver | | (logs/errors/metrics/spans) | |
| +-----+-----+ +-------------+---------------+ |
| | | |
| v v |
| +-------------------------------------------+ |
| | Ingestion Pipeline | |
| | (normalize, enrich, fingerprint, batch) | |
| +--------------------+-----------------------+ |
| v |
| +-------------------------------------------+ |
| | PostgreSQL (partitioned by day) | |
| +--------------------+-----------------------+ |
| v |
| +-------------------------------------------+ |
| | Web UI (HTMX + SSE + Tailwind/DaisyUI) | |
| +-------------------------------------------+ |
+-------------------------------------------------+
Single binary. PostgreSQL is the only external dependency.
This work is licensed under CC BY-NC-SA 4.0.
You are free to:
- Share and adapt this software for non-commercial purposes
Under the following terms:
- Attribution — Credit to Michal Lipinski with a link to this repository
- NonCommercial — No commercial use without a separate license agreement
- ShareAlike — Derivatives must use the same license
For commercial licensing inquiries, contact the author.