Evenfire is an intelligence layer platform for running LLM agents that take real actions, on your infrastructure
TypeScript
import { registry } from './metrics'
import { startCoordinatorHealthServer } from './workflow/coordinatorHttpServer'
import { runWorkflowFromEnvironment } from './workflow/sdkRuntime'
let healthPhase = 'initializing'
async function main(): Promise<void> {
console.log('[Coordinator] Starting workflow coordinator')
const healthServer = startCoordinatorHealthServer({
getPhase: () => healthPhase,
metricsRegistry: registry,
})
let exitCode = 1
try {
healthPhase = 'running'
const result = await runWorkflowFromEnvironment()
healthPhase = result.workflowPhase
exitCode = result.exitCode
console.log(`[Coordinator] Execution complete: ${result.workflowPhase}`)
if (result.failureReason) {
console.error(`[Coordinator] Failure reason: ${result.failureReason}`)
}
} catch (err) {
healthPhase = 'failed'
console.error('[Coordinator] Fatal runtime error:', err)
} finally {
await new Promise<void>(resolve => {
healthServer.close(() => resolve())
setTimeout(resolve, 2000)
})
}
process.exit(exitCode)
}
if (require.main === module) {
main().catch(err => {
console.error('[Coordinator] Fatal startup error:', err)
process.exit(1)
})
}