Integrations catalog
Orbit Pulse is planned to get an Integrations section where a team connects its own databases and services. This page is for administrators who want to know what is coming and what to use in the meantime.
What is planned Planned
Section titled “What is planned ”- A catalog with search and categories: databases, caches, queues and streams, web servers, clouds and custom metrics.
- A Connect wizard with five steps. You say where the database runs and choose or install an agent. You create the monitoring role from ready SQL, decide where the password lives and check the connection. The check reports problems in plain words, without secrets.
- Agents in your network that only connect outward. An agent opens an outgoing HTTPS connection to Pulse and nothing else. Pulse does not reach into your private network.
- Secrets beside the agent. By default the password stays on the agent’s host. Pulse stores only a reference to it, such as an environment variable, a file, a Docker or Kubernetes secret, or Vault.
- Read-only access with minimal privileges. Each engine gets ready SQL or commands for a monitoring role. A superuser is never needed.
- Values stay on your host. The agent normalizes query text before sending it. Separate switches turn query text and plans off.
- Open standards. OTLP, Prometheus and OpenMetrics, and metrics from the OpenTelemetry Collector. A team that already has exporters can connect without the Pulse agent.
Project administrators would create and change integrations. Every project member would see them, and only administrators would see agents.
Engines in the plan Planned
Section titled “Engines in the plan ”| Order | Engines |
|---|---|
| First | PostgreSQL for team databases: several instances, all databases of an instance, managed services |
| Next | Prometheus and OpenMetrics, the OpenTelemetry Collector, then MySQL and MariaDB |
| Then | Redis and Valkey, MongoDB, ClickHouse |
| Later | Queues and streams (RabbitMQ, Kafka, NATS) and agentless cloud integrations |
The order follows the plan and is not a schedule. Basic monitoring would give every engine metrics, health, a dashboard and monitors. Query-level monitoring (queries, plans, sessions, locks) is planned for PostgreSQL, MySQL and MariaDB, and in part for MongoDB and ClickHouse.
What you can use today
Section titled “What you can use today”- Send telemetry yourself. Logs, traces and metrics go to Pulse over OpenTelemetry (OTLP/HTTP), and logs and metrics also as JSON. See Send telemetry with OpenTelemetry and Metrics, dashboards and SLOs.
- Check from outside. Scheduled HTTP checks with assertions on status, latency and body, and monitors that notify Orbit people and call signed webhooks. See Monitors and alerts.
- Link to Forge. Pulse reads deployments and CI results from Forge. See Deployments.
Database monitoring exists today only for Orbit’s own PostgreSQL database. Connecting your own database is planned; see Monitor your PostgreSQL databases.