Control-M alternative
Enterprise job scheduling without a scheduler database or a licence that counts jobs.
Dagu is a Control-M alternative for corporate IT teams whose jobs already run on Linux and Windows servers as scripts, binaries, or containers. Folders become YAML in git, the scheduler database disappears, the licence counts Dagu servers, and Dagu Console monitors every server. This page also covers the migration path.
# nightly-closing.yaml
schedule: "0 2 * * *"
# Business-day calendar
calendar: jp-banking
# Run on the business server without a resident agent
ssh:
user: batch
host: app01.corp.local
key: ~/.ssh/batch_key
steps:
- id: closing
run: /opt/batch/run-closing.sh
retry_policy:
limit: 2
interval_sec: 300
- id: transfer
run: /opt/batch/send-to-core.sh
depends: [closing]
# Recovery job
handler_on:
failure:
run: /opt/batch/recovery.sh
mail_on:
failure: true
Execution order
One binary, no scheduler database to operate
Licensed per Dagu server, with unlimited workers
Folders as reviewable YAML in version control
Runs on-premises and inside closed networks
At a glance
Control-M vs. Dagu for corporate batch
One binary with file-backed state; queues and workers are optional.
Enterprise Manager, Servers, per-host Agents, and a database.
Declarative YAML reviewed in git.
Planning GUI, with JSON definitions through the Automation API.
Free for unlimited servers; paid plans licensed per Dagu server with unlimited workers.
Commercial, by quote, in units such as tasks.
Active and standby schedulers on shared storage, with automatic failover.
Primary and secondary servers with database replication; automatic failover depends on the database option.
Enterprise agreements with first-response targets.
Vendor support contract.
In depth
How Dagu does it
Keep the job flow, drop the infrastructure
Control-M runs as an Enterprise Manager, one or more Servers, an Agent on each executing host, and a database behind them. Dagu keeps the same scheduling concepts inside a single process that stores its state in files.
- Folders become DAGs, jobs become steps, and in and out conditions inside a folder become depends.
- Calendars map to business calendars, and cyclic jobs map to a schedule or a repeat_policy.
- No scheduler database to size, back up, or patch.
Migrate one folder, not the whole estate
Control-M's own export gives you the input, and you move one folder at a time. If you want a migration tool for Control-M definitions, request it at dagu.sh/contact.
- Export folder definitions as JSON or XML, then map them: job to step, condition to depends, failure actions to handler_on.failure.
- Convert with a script you rerun, or let an AI client connected over MCP write each DAG and check it against the server before saving.
- Run dagu validate on every generated file, then run the pilot folder in parallel with Control-M before cutting over.
Definitions can live in git from the start
Control-M offers jobs as code through its Automation API, alongside the planning GUI. In Dagu the YAML file is the definition, so it can live in git and changes can be reviewed there.
- Every job flow is a YAML file, so a team that keeps them in git gets reviewable diffs.
- dagu validate checks generated and hand-written definitions in CI before they reach the scheduler.
- The web UI edits the same YAML, and handles operations: start runs with parameters, watch logs, retry, and suspend schedules.
Reach existing servers without an agent on each one
Batch work usually lives on servers that already exist. Dagu can drive them over SSH, or run a worker process where a persistent executor is the better fit.
- Use the SSH executor to run commands on Linux and UNIX hosts with no resident agent.
- Workers connect outbound over mTLS-protected gRPC, so segmented networks need no inbound ports.
- Match steps to labelled workers with worker_selector. Workers are not licensed.
A cost structure that does not grow with job count
Control-M is sold by quote, in units such as tasks. Community self-host is free with unlimited servers and workers, and the licensed tier has a listed price per Dagu server.
- Community covers the full orchestration engine, web UI, cron scheduling, and the Docker, SSH, and HTTP executors.
- Licensed self-host starts at $500 per year for three server licenses and adds SSO, RBAC, audit logging, and incident routing.
- Adding jobs or executing hosts changes no licence count.
Deadlines and alerts across every server
Dagu Cloud is the control plane for your Dagu servers. Dagu Console watches every server's jobs and raises an alert when one needs a person.
- Give a workflow a deadline, and get an alert when it has not succeeded by then, or when a scheduled run did not start.
- Alerts go by email, Slack, Microsoft Teams, PagerDuty, or webhook, and inside a closed network the control plane can run in your network under an Enterprise agreement.
- Move files with SFTP, S3, and rclone steps, and call services with HTTP, SQL, and Kubernetes steps.
FAQ
Practical questions before adopting Dagu
Can Dagu replace Control-M?
Yes. Folders, jobs, conditions, calendars, cyclic jobs, and failure actions map to Dagu, and Dagu Console adds deadline and late-start alerts across every server. If you rely on something Dagu does not cover yet, request it at dagu.sh/contact.
How do existing definitions move over?
Export the folder definitions as JSON or XML and convert them to DAG files with a script, or let an AI client connected over MCP write each DAG and check it against the server. dagu validate checks every file. If you want a migration tool for Control-M definitions, request it at dagu.sh/contact.
What replaces in and out conditions?
Inside one workflow, depends expresses the order. Between workflows, one calls another with a dag.run step, and a wait step can hold a workflow until a file appears or an HTTP endpoint is ready.
What replaces cyclic jobs?
A cron schedule covers intervals of a minute or more. Use repeat_policy on a step to repeat it within a run until a condition is met.
Does it work in a closed network?
Yes. Dagu is a self-contained binary with no required external database or broker. For deadline and late-start alerts inside a closed network, the control plane behind Dagu Console can be deployed in your network under an Enterprise agreement.
Next step
Start with one workflow.
Install Dagu, move one fragile script or agent task into YAML, and decide from a real run history.