ISSUE 002 / SEPTEMBER 2026
FROM JAMS
THE QUIET HOURS
JAMS Software’s monthly newsletter.
This month is the best month of proof we have had in a while. Kent State rebuilt 182 Banner job chains and cut over in 26 days with no contingency. JAMS Incident Management is public. And 7.9.2 quietly fixed something that has been annoying anyone running a lot of File Watches.
BS
Bobby Schmidt
VP of Marketing, JAMS Software
01 / CUSTOMER STORY
182 Banner job chains, 26 days, zero contingency
It does everything the old scheduler did, and it gives us more control over how things run.
Amanda Kelley-Hurm, Senior Manager, Application Administration Support, Kent State University
Kent State University’s old scheduler contract expired on June 28. No extension, no fallback. Everything had to be running on JAMS Scheduler by that date. They rebuilt all 182 Banner job chains and cut over on the day the contract ran out. Twenty-six days from kickoff, no contingency days built in, and no failed or missed runs since. Promoting a chain to production used to take up to an hour. It now takes minutes.
02 / WHAT SHIPPED
JAMS Scheduler 7.9.2
The part that matters
File Watches are now monitored on the backend instead of one thread per watch. If you run a lot of them, this is the release that stops you worrying about thread exhaustion. No configuration changes needed.
Released September 3. Two things worth your attention. System API Tokens. You can now issue tokens for service accounts and integrations, so scripts call the JAMS REST API without an interactive sign-in. Tokens are managed from a new API Tokens page in the Web Client, and each one inherits the permissions of the user it is bound to. No new permission model to learn. Also in 7.9.2: Sequence Jobs get a Sequence tab in the Monitor with a real-time task tree, so you can skip, hold, release, and retry individual tasks without leaving the Monitor.
03 / NOW GENERALLY AVAILABLE
JAMS Incident Management is public
On-call scheduling, alerting, escalation, and incident management. Voice, SMS, push, email, Microsoft Teams, and Slack. Escalation policies, on-call rotations, and a full audit trail of who was paged, when, and what happened next. It integrates with your orchestration, monitoring, and service desk tools, and it does not require JAMS Scheduler. If you run both, a failed job can open an incident on its own. Starts free. Paid tiers begin at $6 per user per month, published on the page rather than quoted on a call.
04 / FREE TOOL
JUMP converts your old scheduler for you
JUMP is the JAMS Universal Migration Platform. It is free, it is standalone, and it converts job definitions from your existing scheduler into JAMS Scheduler-ready jobs. Start with a Dry Run, which needs no connection to anything. Point it at your definitions and see what a conversion would produce before you commit to it. Connect it to a JAMS Scheduler instance when you are ready for the real thing. Windows Task Scheduler, Cron, and SQL Server Agent convert fully automatically in most environments. Tidal, Control-M, ActiveBatch, VisualCron, AppWorx, and Airflow convert with assistance.
05 / COME SAY HELLO
JAMS Scheduler at SQL Saturday Boston, October 3
We are sponsoring SQL Saturday Boston on Saturday, October 3. Louis Diaz is presenting SQL Server’s Alerting Blind Spot: Automating Jobs, Deployments, and Incident Response, 12:00 to 12:45 in Room 3. If you are in the area, we would love to see you.
06 / ONE MORE THING
The Golden Pager is up for grabs in Boston
level_12: still not kind.
If you missed it last month: Dependency Hell is our browser game about the part of the job nobody sees. A queue of jobs, a couple of lanes, and a window that closes at 6 AM. Place them by hand, then watch the same jobs run again triggered by dependency instead of by the clock. The gap between the two times is the entire argument. We are bringing it to SQL Saturday Boston on October 3, and the Golden Pager goes to whoever posts the best time on the day. Worth a few practice runs. Twelve nights, free, no signup, works on your phone.
Worth reading
Fewer than 5 percent of IBM i shops run IBM i only. Which means the hard problem was never IBM i scheduling. It is what happens when an IBM i job has to hand off to something that is not IBM i.
The launch write-up for JAMS Incident Management, including why we think per-seat pricing in this category stopped making sense a while ago.
What else we’re reading
A permissions change caused a Bot Management configuration file to double in size. The code parsing it was not expecting a file that big, panicked, and took core traffic routing down for roughly two hours. Cloudflare’s write-up is unusually direct about the whole chain. The lesson is not about bot management, it is that the thing which breaks you is often a file, not a server.
blog.cloudflare.com
On September 3, ChatGPT, Claude, and Grok all failed at roughly the same time, with Azure showing elevated errors the same morning. None of the three confirmed a shared root cause. The useful part is the reminder that vendor diversity on paper is not always vendor diversity underneath.
axios.com
Lex Neva has been rounding up the week in reliability engineering for over five hundred issues, and the outage reports section alone is worth the subscription. If you like the postmortems we keep linking here, this is where a lot of them surface first.
sreweekly.com
You’re receiving this because you’re a JAMS Scheduler customer or asked to hear from us. One email a month, and the unsubscribe link works on the first click.
Get The Quiet Hours in your inbox
One email a month. Send me The Quiet Hours, and nothing else.
