Migrate Legacy schedulers

Keep the scheduler. Move the data jobs.

Control-M, AutoSys, Tidal, IBM Workload Scheduler and plain cron run thousands of jobs in most estates, and many of those jobs are data jobs: extracts, loads, file transfers, scripts that stitch them together. The scheduler sees a job succeed or fail; it does not see the rows. This path keeps the scheduler for what it does well and moves the data jobs into Lima pipelines, which the scheduler triggers and hears back from. Each job gains checks, exception owners, a run window with a deadline and a reconciliation report. Estates on Airflow 2, which reached end of life on April 22, 2026 [6], take the same path onto Airflow 3.

1InventoryList the scheduler's datajobs: extracts, loads,transfers, scripts; theirdependencies andcalendars.2Rebuild as pipelinesEach data job becomes aLima pipeline with checks;scripts become operatorsor Python.3Wire the triggerThe scheduler starts theLima job through theexternal-scheduler triggerand receives its status.4Run side by sideOld and new run together;rows and totals arecompared per table.5Cut overThe scheduler's job nowpoints at Lima; the oldscript is disabled.
The scheduler stays in charge of the estate's calendar. Lima runs the data jobs inside it and reports back.
5
steps from inventory to cut-over
offer
100
concurrent SAP jobs per Control-M agent
vendor docs [1]
Apr 14, 2026
AutoSys 24.2 released
dated source [3]
Apr 24, 2026
IBM Workload Scheduler 10.2.7 released
dated source [5]
Apr 22, 2026
Airflow 2 end of life
dated source [6]
What changes, and when

Nothing is ending; something is missing.

These schedulers are current products with recent releases [3][5]. The reason to move data jobs is what a scheduler cannot see: the rows. A job that finishes on time with 3 rejected purchase-order lines looks green to the scheduler. In Lima those 3 lines are exception records with owners, the totals reconcile, and the report says so.

SchedulerStatus on the check dateSource
BMC Control-MCurrent; Control-M SaaS and on-premise; Control-M for SAP certified; SLA Management add-on for deadline chains.[1][2]
Broadcom AutoSys Workload AutomationCurrent; release 24.2 on April 14, 2026.[3]
Tidal Automation (Redwood)Current; sold by Redwood since its 2023 acquisition of Tidal Software.[4]
IBM Workload SchedulerCurrent; release 10.2.7 on April 24, 2026.[5]
Apache Airflow 2End of life on April 22, 2026; Airflow 3 is current.[6]
cronEverywhere, forever, and blind to rows.—
What moves where

Each thing you have today, and what it becomes.

The conversion is a mapping, and the mapping is short. What has no equivalent is listed too.

TodayIn LimaWhat to know
A data job (extract, load, transfer, script)A Lima pipelineSame sources and targets, plus checks, exception owners and a report.
Job dependency or conditionDependency triggerA pipeline starts when the ones it depends on finish; the scheduler's condition can remain as the outer trigger.
CalendarBusiness calendarWorking days, cut-offs and run windows with deadlines.
SLA chain (Control-M SLA Management)Run window with a deadline clock inside the jobThe chain says whether the job finished by the deadline; Lima says what happened inside, and pauses the clock when a check trips.
File watcherEvent triggerArrival of a file on SFTP or object storage starts the pipeline.
Alert to an operatorException routed to its ownerA bad row goes to the buyer or planner who owns it, with a deadline; an operational failure still alerts operations.
Return codeStatus back to the schedulerLima reports success, held (paused on a check) or failed, with the report link.
RetryResume from checkpointThe run continues from where it stopped and replays in order; nothing is applied twice.
How the conversion runs

Five steps, each with an acceptance record.

Inventory (week 1, free)We read your scheduler job definitions (exports or JIL files) and count the jobs by pattern: simple copies, lookups and joins, incremental logic, scripts, change capture, write-back. Each gets an effort band and a flag for anything Lima does differently.
Convert the first ten to twenty (weeks 2–4, fixed price)Each job becomes a Lima pipeline with the same sources and targets. Scripts become Python operators. Reject handling becomes exception records with owners.
Run both, compare the resultsThe old and the new run side by side for an agreed number of cycles. Rows, control totals and sample values are compared per table, and the differences are explained or fixed.
Cut over, one job at a timeWhen a job reconciles for the agreed cycles, its schedule moves to Lima and the old one is disabled. The reconciliation report is the acceptance record.
RetireWhen the last job has moved, the old server or subscription is switched off. The inventory shows what is left at every step.
The assessment

One week to know what you have.

The assessment reads exported job definitions (JIL for AutoSys, XML for Control-M, exports for Tidal and IWS, crontabs) and needs no access to production data.

What it readsWhat it reports
Job definitionsWhich jobs are data jobs, by type: extract, load, transfer, script.
Dependencies and calendarsThe chains the data jobs sit in and the deadlines implied.
ScriptsShell, SQL and Python by size, with the conversion effort.
Effort bandsPer batch of ten to twenty jobs.
The offer

Free to look, fixed price to move.

What stays

What you keep.

The scheduler stays. Its operators see the same chains and the same green and red; Lima adds a status of 'held' when a check trips, with the report link.

SAP batch jobs stay in Control-M for SAP or its equivalent; Lima's SAP work is data extraction and write-back through released interfaces, which is a different job.

Terms on this page

Plain meanings.

Workload automationScheduling and running jobs across an enterprise with dependencies, calendars and monitoring.
JILAutoSys's job definition language.
External scheduler triggerLima's trigger mode in which an enterprise scheduler starts a Lima job and receives its status.
Run windowLima's term for the period in which a load must run and finish, with a deadline clock that pauses when a check trips.
CheckpointThe saved position from which an interrupted run resumes.
Questions

Asked before a migration.

Why move the jobs at all if the scheduler is current?

For what the scheduler cannot see: rows, totals, exceptions and who owns them. The job looks the same to operations; the data team gets the report.

Will the operations team lose visibility?

No. The scheduler still starts the job and receives success, held or failed, with a link to Lima's report.

We are on Airflow 2. Is that the same path?

Yes. Airflow 2 reached end of life on April 22, 2026 [6]. The DAGs move to Airflow 3 with Lima tasks inside them, following the same five steps.

Start with the inventory.

One week, no charge. You get the count by pattern, the exposure flags and the effort bands, whatever you decide next.