Skip to content

Know when your services need attention.

AlertBird is a self-hosted tool for service health and alerting in Kubernetes. It turns OpenTelemetry data into a map of your services, recent health history and alerts sent from your own infrastructure.

Storefront sends requests to checkout. Checkout calls payments, which has an elevated error rate, and postgres.

checkout to payments

requests/min
240
p95
180 ms
errors
1.2%

Error rate above 1%

Sent to Slack

See how your services depend on each other.

AlertBird builds a service map from observed traffic between instrumented services.

Open a service or connection to inspect request volume, errors, latency and HTTP status codes. Missing or stale telemetry stays visible.

Explore service health

checkout to payments

Elevated errors

requests/min
240
p95 latency
180 ms
errors
1.2%

Error rate

Last hour

Error rate over one hourErrors rise above the one percent rule threshold toward the end of the hour.

HTTP responses

200 98.8%

500 1.2%

Error rate above 1%

Sent to Slack

An error-rate line crosses its threshold beside an alert bell and a notification sent to Slack.

Send alerts from your own infrastructure.

Set rules for error rates, latency, selected metrics, matching logs or stale telemetry.

Send notifications directly to Slack, Teams, email, PagerDuty or webhooks. Follow delivery activity and recovery in AlertBird.

Set up alerts

Keep a recent view of service health.

AlertBird keeps compact health summaries and selected diagnostic evidence, with a default monitoring history of seven days.

Use PostgreSQL in your cluster or connect a database you already run.

What AlertBird keeps
Seven thin chart leaves form a compact folio of recent health history, from seven days ago to now.

Use the OpenTelemetry setup that fits your cluster.

Bring your existing collection setup, or start with the one included with AlertBird.

Starting with a Kubernetes cluster?

Use the included upstream Collector setup. Supported Node.js and Python automatic instrumentation is opt-in.

Use included collection

How it fits together.

OpenTelemetry collects and forwards the data. AlertBird turns it into a view of service health and evaluates your alert rules.

  1. OpenTelemetry

    Instrumentation
    Collection and export

  2. AlertBird

    Health · error groups · rules
    Dashboard and alert delivery

  3. PostgreSQL

    Recent history
    Configuration and delivery state

Read the architecture

Help shape AlertBird.

AlertBird is an early project. If you try it, we’d like to hear what worked and what got in your way.

Report a problem, suggest an improvement or help with the code and documentation.

The repository is private while public source access and licensing are prepared.

A few things to know.

Can I use my existing Collector?

Yes. Send supported OpenTelemetry data from your existing Collector or SDKs to AlertBird’s endpoint. Included Kubernetes collection is another option.

Does AlertBird keep every log and trace?

It keeps compact health history and selected diagnostic evidence. A searchable archive of every raw log and trace is outside the current scope. Use a separate backend when you need that archive.

Do I need an AlertBird cloud account?

The self-hosted core runs in your infrastructure and sends alerts from its own worker. You provide the Kubernetes cluster, database and credentials for your alert destinations.

What do I need to install it?

The quickstart covers source-built images, credentials, TLS, collection and PostgreSQL. You can use the included database or an external PostgreSQL instance. Repository access is currently required while the public source release is prepared.

Start with your services.

The quickstart walks through a self-hosted installation.

Get started