What's New
We ship improvements continuously, and this page is where they land. It's updated as we release new features, enhancements, and fixes — so check back to see what's changed.
🛡️ Choose which JAMS actions your team can take
Updated — September 30, 2026
Admins can now decide which remediation actions a JAMS Connection offers. When a JAMS job fails, recipients are offered actions like Restart, Continue and Cancel, and until now that was everything that applied. You asked for a way to stop some of them being used. The request we heard most was "I would never want a job cancelled, especially a Sequence or a Workflow." Each Connection now has six checkboxes: Restart job, Retry step in workflow, Continue sequence / workflow, Cancel job, Cancel sequence and Cancel workflow. Untick the ones you don't want offered.
How to set it up: go to JAMS Scheduler, click Edit on a Connection, untick actions under Actions recipients can take, and save. Only Admins can change it. Everything stays allowed until you do, so nothing changes for your team unless you choose to. See Choosing which actions your team can take.
It applies everywhere at once. An unticked action is removed from the alert on voice, SMS, WhatsApp, Teams, Slack and push, and the menu is renumbered so there are no gaps. Acknowledge is always offered, so someone can always take ownership of an alert. There's nothing to change in JAMS itself.
It takes effect immediately, even on alerts already sent. If somebody chooses an action you've since switched off, it's refused and doesn't run. The call doesn't just end: the menu is read again with only the choices still allowed, so they can pick another action or acknowledge. On SMS and WhatsApp the reply lists the digits they can still send. Every refusal is recorded in the audit log.
💬 WhatsApp alerts, and planned maintenance without muting anything
Updated — September 30, 2026
WhatsApp is a notification channel. Alerts arrive in WhatsApp and responders reply with a digit to acknowledge — or, for a JAMS Scheduler failure, to restart, continue or cancel the job. It matters most where SMS lets you down: in India the carrier route rewrites the sender ID so replies can never come back, and in Peru two-way SMS is unavailable at all. WhatsApp is the only channel that carries a working reply in those places. Each person verifies their own handset from their Settings page, and WhatsApp messages draw on the same allowance as SMS rather than a separate one. See WhatsApp.
Pause alerting for planned maintenance. Patch a server, run a migration, take a system down for the night — and stop the alerts it was always going to produce, without touching a single flow. Pause everything now for a set length of time, schedule one for later, or set a repeating window (every Sunday 02:00–04:00, say). You can scope a pause to specific flows, so the rest of your alerting carries on. Suppressed alerts are dropped, not deferred — nothing floods in when the window closes — and they cost you nothing from any allowance. The bar shows how many were suppressed, because a pause that looks identical to a quiet night is worse than no pause at all. See Pausing alerts for maintenance.
A maintenance window never resurrects a flow you had switched off. Ending a pause restores exactly the state you were in, including flows that were already inactive on purpose. And every window has an end time by arithmetic — there is no open-ended option, because in practice those become "nobody remembered to turn it back on".
Coverage hours you can actually express. On-call coverage is now a list of windows rather than a day-of-week mask, so "Monday 09:00 straight through to Friday 17:00" is one window instead of something the old model could not represent at all. Overnight windows and daylight-saving transitions are handled properly.
PagerDuty imports land on the right hours. Three fixes worth naming, because each produced a schedule that looked plausible and was wrong: imported shifts were arriving five hours out; a schedule's coverage hours displayed as 24/7 in the editor and were then destroyed on the next save; and overrides were only read a fortnight ahead, so anything further out was silently missed. Imports now read a year ahead. See Migrating from PagerDuty.
Acknowledgements say who, and how. A Teams or Slack card now reads "JAMS restart by Kevin Casey via WhatsApp reply" rather than naming an internal action code. It names the person where we can identify them, the channel the response actually arrived on, and — for a JAMS remediation — the action they chose, in the words they chose it in. The card also no longer prints the alert twice when the title and the message are the same sentence.
You get told when a channel cannot reach someone. Add a recipient to a flow, a JAMS Connection, Azure Monitor, ServiceNow, Zendesk or a Windows service monitor, and anyone who cannot be reached on a channel you ticked is named on the spot — no phone number on file, or a WhatsApp handset they have not verified yet. The channel stays on the flow either way, so it starts working the moment they fix it.
Notify everyone on a flow at once. A flow that lists several people can now page them simultaneously instead of walking the list one at a time, without needing an escalation policy.
Integrations are in alphabetical order in the sidebar, which matters now there are ten of them.
Smaller things. Saved browser passwords stopped autofilling into integration secret fields. API keys show a webhook URL you can actually copy, and a Last Used that reflects reality. The usage meters on the billing page show only the allowances that are actually metered. Windows service monitoring and JAMS service monitoring no longer show each other's rows, and a service ticked on the Windows page now reaches the dashboard card it belongs to. The JAMS connector works out your API path prefix instead of assuming one. A master admin can place a tenant on a trial directly.
🔌 Monitor any Windows service, and migrate off PagerDuty
Updated — August 28, 2026
Windows Service Monitoring. Get paged when any Windows service stops — a SQL Server instance, a vendor agent, your own background worker — on any Windows server. Add the server, run the installer, and use Discover services to list what's actually installed on it rather than typing names in. No inbound firewall rules and no credentials: the Agent reports outbound and runs as LocalSystem. If a server already runs the JAMS Agent, there's nothing extra to install. See Windows Service Monitoring.
A silent server is no longer reported as healthy. Previously a powered-off machine showed every service as Running, because nothing had told us they'd stopped. Now an Agent that goes quiet raises its own alert saying exactly that, and its services show Unknown rather than a state we can't actually see. Service labels also name the machine they're on, so two servers running the same service stay distinguishable.
We now tell you when a monitor can page nobody. Monitoring switched on, services ticked, and no alert flow chosen — or the flow was later deleted, deactivated, or left with no reachable recipient. Any of those and the page says so, instead of looking healthy until an incident.
Migrating from PagerDuty. Bring across people, teams, schedules (both of PagerDuty's schedule types), rotations, escalation policies and overrides. You get a full report first — what would be created, the resulting coverage, and anything that can't be represented faithfully — and nothing is written until you press Import. Nobody is emailed: imported people are invited individually, from the Users page, when you're ready to cut over. See Migrating from PagerDuty.
JAMS service monitoring and queue depth. Watch the JAMS Executor, Scheduler and Server services, and get paged when one stops — with an Attempt to start button right on the alert and the dashboard. MSMQ queue depth is monitored too, with a threshold you set, sustained-breach detection so a brief spike doesn't page anyone, and a recovery notification when it clears. See JAMS Scheduler.
Per-user pricing, with SMS and voice metered separately. Plans are now priced per user, and seats are something you buy rather than a count of who happens to be in the account — so inviting or removing somebody never changes your bill. SMS and voice have their own allowances, which means running out of texts doesn't stop calls, and neither stops email, push, Teams or Slack. Top-ups are available for both. See Plans & Billing.
More than one person on call at once. A schedule can now be set so that every rotation covering a moment pages, simultaneously — for a deliberate handover overlap, or a shared pager — instead of the top rotation winning and the rest acting as fallbacks. Uncovered windows are also surfaced everywhere now: on the calendar, in the timeline, and on the dashboard. See Rotations and coverage and Coverage gaps.
Who's on call, on the dashboard. Today at a glance now lists who currently holds each schedule's pager, marks you if it's you, and names any schedule with nobody on call rather than hiding it in a count.
One alert, one row. The dashboard and history pages now collapse a multi-channel alert into a single expandable row, so an alert that went out by SMS, call and email reads as one event. History also gained an Ack Time column and collapsible filters.
AI Helper works with Anthropic too. Paste an Anthropic key instead of an OpenAI one and it's used automatically — the vendor is chosen from the key itself, with nothing to configure. See AI Helper.
Migrating from Opsgenie. A read-only check that reports exactly what would come across from your Opsgenie account. Worth running on its own — it finds problems in a current setup whether or not you migrate. Opsgenie closes on 5 April 2027. See Migrating from Opsgenie.
Clearer JAMS alerts and remediation. A failure alert now names the sequence or workflow the job belongs to, not just the job. Choosing Retry on a workflow tells the caller plainly that it starts a new job rather than resuming the old one, and cancelling a run no longer reports it as having finished successfully. JAMS's newer API is supported as well, with setup problems reported as something you can act on instead of a bare error.
Dashboard tidying. New cards for Windows Service Health and queue depth (with your alert threshold drawn on the graph), and the plan-usage meters are two lines each instead of five.
Trials are gentler. A trial is now metered per channel rather than ending the moment you hit a single cap, and an expired trial continues on Free instead of locking you out. The billing page shows trial status with days remaining, and there's an Enterprise option on it.
Privacy and security. Marketing analytics no longer run anywhere inside the signed-in app — only on public marketing and sign-up pages, with sign-in and other forms explicitly excluded. You can also anonymise one person's personal details in the audit trail on request, keeping the trail itself intact. And we fixed a flaw in how a caller's IP address was determined, which had made login rate-limiting bypassable and audit-trail IPs forgeable.
AI Helper answers read properly again. Answers grounded in this Help Center had been losing their numbered lists, and long articles were being cut short before the model saw them. Both fixed, so answers are better and formatted the way they were written.
⚡ A faster flow builder + JAMS multi-environment routing
Updated — July 24, 2026
Pick your source first. Creating a flow from New Alert Flow now starts with choosing what to monitor, then shows just the settings for that source — name, description, custom alert message, and configuration together in one step. Fewer clicks, less back-and-forth.
JAMS Scheduler and Azure Monitor are right in the flow builder. If your plan includes them, they show up as choices under New Alert Flow and take you straight to setup. After you save, a View Flows button drops you on your flow list.
Manage integration flows from the Flows list. JAMS and Azure Monitor flows now have Edit (which opens their setup form) and Delete right alongside your other flows.
JAMS: one alert path per environment. Running JAMS in more than one place (say Prod and Dev)? Install a connector in each, then bind a flow to the environment it belongs to — so a restart, continue, cancel, or acknowledge only ever runs against the right server. See JAMS Scheduler.
Clearer save confirmations. Saving now shows a brief on-screen confirmation, so you always know it worked.
🚀 Welcome to JAMS Incident Management
Launched — July 9, 2026
JAMS Incident Management makes sure the right person hears about the alert that matters — and can act on it from wherever they are. Here's what the platform can do:
Alerts on every channel Reach your team by SMS, phone call, email, push notification, Microsoft Teams, and Slack — choose the channels each person prefers, per alert.
Never miss a critical alert Sequential escalation, on-call rotations, and team groups automatically move an unacknowledged alert to the next responder until someone owns it.
Acknowledge and act from anywhere Reply to a text, tap an email or push notification, press a key or speak on a call, or click in Teams or Slack — no need to open the app. For JAMS Scheduler jobs, you can restart, continue, or cancel a failed job right from the alert.
Connect the tools you already use Two-way integrations with JAMS Scheduler, Azure Monitor, ServiceNow, and Zendesk — acknowledge or resolve here and the status syncs back to the source automatically.
Build alert flows in minutes A visual flow builder — or let the built-in AI Helper draft a flow for you from a plain-English description.
Some capabilities depend on your plan or add-ons — see Plans & Billing for what's included in yours.
Questions or ideas for what we should build next? We'd love to hear them. We're just getting started.