Composite Monitor

The Composite monitor rolls up the status of several existing monitors into a single derived monitor. It does not perform any network checks of its own — it reads the latest status of the referenced monitors from updu's storage and applies a quorum rule.

Use it to express service-level health on top of probe-level monitors (for example: "checkout is healthy if at least 2 of the 3 web nodes are up").

Configuration Options

Basic Settings

• Name: A descriptive name for your composite (e.g. Checkout (any node up)).
• Group: Optional group assignment for organizing monitors.
• Interval (seconds): How frequently updu re-evaluates the composite. Composite checks are inexpensive — short intervals are fine.

Composite Specific Settings

• Monitor IDs: The set of underlying monitor IDs to evaluate. Required.
• Mode: How updu decides whether the composite is up:
  • all_up — every referenced monitor must be up.
  • any_up — at least one referenced monitor must be up.
  • quorum — at least Quorum of the referenced monitors must be up.
• Quorum: Required when mode is quorum. The minimum number of underlying monitors that must be up.

If a referenced monitor has no recorded status yet (for example because it has never run), it is treated as not-up for quorum purposes.

Example Use Cases

• Cluster health: Express "API cluster healthy" as quorum=2 over three regional probes — one slow region won't trigger a page.
• Failover paths: Use any_up to roll up "primary or backup" connectivity into a single rollup status.
• Critical-path gating: Use all_up to gate a status page section behind every dependency being up.