kestra-io/kestra · critical · RuntimeException

{} not started in time

Error message

{} not started in time

What it means

Thrown by StandAloneRunner when one or more embedded services (Controller, Worker, Scheduler, Indexer, SystemWorker) fail to reach RUNNING/MAINTENANCE state within the configured startup timeout (default 1 minute via ServerConfig.Standalone.Running.timeout). The message lists each lagging service with its simple class name and current ServiceState. It is a bare RuntimeException wrapping a ConditionTimeoutException from Awaitility.

Source

Thrown at cli/src/main/java/io/kestra/cli/StandAloneRunner.java:111

            poolExecutor.execute(scheduler);
            servers.add(scheduler);
        }

        if (indexerEnabled) {
            Indexer indexer = indexerProvider.get();
            poolExecutor.execute(indexer);
            servers.add(indexer);
        }

        // start the embedded SystemWorker (always present in STANDALONE mode)
        SystemWorker systemWorker = systemWorkerProvider.get();
        poolExecutor.execute(systemWorker::start);
        servers.add(systemWorker);

        try {
            Await.await().atMost(getRunningTimeout()).until(() -> servers.stream().allMatch(StandAloneRunner::isStarted));
        } catch (ConditionTimeoutException e) {
            throw new RuntimeException(
                servers.stream().filter(s -> !isStarted(s))
                    .map(s -> s.getClass().getSimpleName() + " (state: " + s.getState() + ")")
                    .toList() + " not started in time"
            );
        }
    }

    /**
     * A service is only considered started once it reached RUNNING (or MAINTENANCE).
     */
    private static boolean isStarted(Service service) {
        Service.ServiceState state = service.getState();
        return Service.ServiceState.RUNNING == state || Service.ServiceState.MAINTENANCE == state;
    }

    private Duration getRunningTimeout() {
        return Optional.ofNullable(serverConfig.standalone())
            .map(ServerConfig.Standalone::running)

View on GitHub (pinned to 823fada927)

Solutions

1. Read the listed service(s) and their ServiceState in the message — that names which component is stuck.
2. Check the logs above this exception for the underlying cause (DB connection refused, port bind failure, plugin load error).
3. Verify all required backends (database, queue, search) are reachable and credentials are valid.
4. Increase kestra.server.standalone.running.timeout if the environment is legitimately slow (e.g. cold H2/Postgres).
5. Ensure no port conflict (default 8080) and that enough CPU/heap is available for startup.

Example fix

// before: default 1 minute may be too short on cold start
Await.await().atMost(getRunningTimeout()).until(...);
// after: raise the timeout in application.yml
# kestra:
#   server:
#     standalone:
#       running:
#         timeout: PT2M

// Before run(): probe required backends
if (!isDbReachable(datasourceUrl)) { throw new IllegalStateException('DB unreachable'); }

try {
  runner.run();
} catch (RuntimeException e) {
  if (e.getMessage() != null && e.getMessage().endsWith(" not started in time")) {
    log.error('Startup timed out; stuck services: {}', e.getMessage());
    dumpServiceStates();
  }
  throw e;
}

When it happens

Trigger: StandAloneRunner.run() submits all enabled services to the pool, then Await.await().atMost(timeout).until(allMatch(isStarted)) times out; the catch builds a list of services where isStarted is false and throws.

Common situations: Database (JDBC queue / repository) unreachable or slow to connect, Kafka/Elasticsearch backend down, port conflict on the worker/controller, a plugin failing during initialization blocks a service thread, low resources (CPU/memory) on cold start, or the standalone running timeout is set too low.

Related errors

• Invalid flow path
• No log store configured through the application property '%s
• Migration lock is held by another process. Another instance
• Scheme not supported: {}
• Both username and password must be provided if either is pre

AI-assisted analysis of kestra-io/kestra@823fada927 (2026-08-14). Data as JSON: /api/errors/39daa82d0af36731. Report an issue: GitHub.