Observability Config

The ObservabilityConfig object is passed directly to the observability() plugin factory.

transports

Type: ObservabilityTransport[]
Required: Yes

Array of transport instances to forward log entries to. Create Loki transports with createLokiTransport(). Implement ObservabilityTransport directly for custom backends.

import { createLokiTransport } from '@bluecadet/launchpad/observability/transports/loki';

observability({
  transports: [
    createLokiTransport({ url: 'http://loki:3100', defaultLabels: { app: 'kiosk' } }),
  ],
});

include

Type: string[]
Default: ["log:*"]

Event name patterns to forward. Supports * wildcards. An empty array forwards all events.

// Also forward lifecycle events alongside logs
observability({
  transports: [...],
  include: ['log:*', 'command:*', 'workflow:*', 'monitor:*', 'system:*'],
});

exclude

Type: string[]
Default: []

Event name patterns to suppress. Takes precedence over include. Useful for silencing noisy events while keeping a broad include list.

// Forward everything except verbose/debug logs
observability({
  transports: [...],
  exclude: ['log:verbose', 'log:debug'],
});

batch.intervalMs

Type: number
Default: 1000

How often (in milliseconds) to flush the batch buffer to transports, regardless of batch size. Lower values reduce latency; higher values reduce HTTP overhead.

batch.maxEntries

Type: number
Default: 100

Maximum number of log entries to accumulate before a forced flush. When this threshold is hit, the batch is sent immediately without waiting for intervalMs.

buffer.maxBatches

Type: number
Default: 50

Maximum number of failed batches to hold in memory while retrying. When this limit is exceeded, the oldest batch is dropped and observability:buffer:full is emitted.

buffer.maxRetries

Type: number
Default: 3

Maximum number of retry attempts per failed batch. Retries use exponential backoff starting at 1 second. After all retries are exhausted, the batch is dropped and observability:push:dropped is emitted.