Tune how Continuum talks to each legacy agent. Turn Check CRC off for an agent that sends no CRC, and choose whether a Windows, SQL or MCP agent's job values are XML-escaped.
Encrypt the connection to your legacy agents. Turn on Use TLS for Scheduling and the relay schedules the agent's jobs and fetches their output over TLS, checking the agent's certificate.
Address an alert from the job's own properties. An email's recipients and message can name a property set on the job or its workflow, and the notification fills it in.
Running now means running. A job reads Running only once its agent confirms the start, and a start that stalls is flagged Still Attempting Start, with an alert you can act on.
Look back at a workflow's runs across days. The new Workflow History report, and View Workflow History in Processes, list each finished run with its times, job count and failed jobs.
Tag a job by the frequency it runs on. Give a frequency its own tags, and every run built under it carries them too, so a notification can single those runs out.
Give every agent its own properties. Set name and value pairs on any agent, read them in jobs with [[MI.Name]], and update them from scripts with $PROPERTY events.
A legacy agent whose relay has been gone for an hour now reads Offline instead of Unknown, and raises Agent Offline if still down when the relay returns.
A Cancel Schedule event now cancels the workflow's held and waiting jobs, so none are left stuck, and the workflow completes once its running jobs finish.
Request limits now count each signed-in person separately, so one busy user no longer causes "too many requests" errors for everyone else in your organization.
Completion Expression events and event commands on a container job that missed its start time no longer see the reason its sub-workflow could not be built.
A Self Service input default that requesters cannot submit is now refused when the input or button is saved, and Copy leaves such a default out, naming the inputs.
Workflow Selection and Job Selection inputs can now have a default, and a list delimiter containing a comma is judged by the event fields that actually receive it.
A $JOB:ADD or $SCHEDULE:BUILD event whose property list names a property with an unresolved token is now refused, instead of adding a property named with the token's text.
A From Start rerun that runs past its interval is checked against its Latest Runtime Offset at its end, so it no longer runs once more after the cutoff.
Decide who manages agents, pools and relays. The Agents page now checks the Agents, Agent Pools and new Relays permissions, granted in all workspaces, and existing roles keep the access they used.
Give event identities a role from the role itself. A role's Members tab now switches event identities onto and off the role, and saves them with the rest of the role.
Hand a relay over before maintenance. Hand Over moves a relay's work to a standby you choose, and the Relays tab flags any online relay with No Standby so you can add one first.
No more deleting an agent your jobs still need. Continuum now refuses to delete an agent that is online, has jobs on it, or is named by a workflow or unfinished job, and says why.
Remove a relay that is gone for good. Deleting a stopped relay that still holds jobs now offers to end them, so you no longer wait for work that can never finish.
Start OS 2200 jobs in XML. OS 2200 agents now start their jobs in XML by default, and a Job Start Format setting keeps older LSAMs on the non-XML record.
Keep your Find Report from RSJ Output failure criteria. A UNIX - Episys: Find Report from RSJ Output job now takes exit criteria, so you decide which codes finish OK, as in Classic.
IBM i jobs laid out the way you know them. IBM i job forms follow the familiar field order in collapsible sections, suggest common values, and grey out fields your settings do not use.
Decide what a Windows File Arrival exit code means. Windows - File Arrival jobs now take exit criteria, as in Classic, and Corelation jobs get a Use Least Busy Batch Queue switch.
SQL jobs that show only what they use. The six SQL job types sort together, each form shows only the fields your choices need, and a file-system DTExec package takes a Configuration File.
Decide what a WS_FTP Pro exit code means. Windows - WS_FTP Pro jobs now take exit criteria, so a transfer can finish OK or fail on the codes you choose, as in Classic.
An MFT Transfer form that follows your transfer. Fields are grouped by step and appear only when their setting uses them, and a complete endpoint or key list accepts only names it holds.
Offer only the job types you use. Turn single job types of a plugin on or off in the Marketplace, see how many jobs use one first, and existing jobs keep running.
Change every job in a workflow instance at once. Job Actions in a Workflow View row menu holds, releases, restarts, cancels or marks every job whose status allows it, and reports any it skipped.
See every z/OS JCL save in Audit History. Each save from a job's JCL dialog is recorded under the person who made it, with its outcome, and never the JCL itself.
Give a Self Service input a starting value. The input editor now sets a default for text, number, date, choice and list inputs, and checks it against the input's own rules.
Copy a job into another workflow. Copy to Another Workflow puts a saved job into another workflow's unsaved changes and opens it there, ready to review and commit.
Schedule a deployment change for a set day. Replace a live workflow deployment with a new version or rules from a future date; the current one keeps running until then.
Match a job's exit description as text. An Exit Description event now takes text as well as numbers, so Equal To SMA0037 fires on exactly that IBM i message ID.
An active relay whose standby stopped more than 30 seconds ago keeps its legacy agents connected through a brief loss of touch with the platform, instead of stepping down.
After the LSAM Connector restarts, the relay reconnects every legacy machine and resumes their status updates, so legacy jobs run again without restarting the relay.
On the New Relay and Rotate Credentials screens, downloading credentials.json now makes Done available, without also finding and ticking I have stored this secret.
Each page now has one page heading, the page name in the header, so screen readers announce where you are, and duplicate titles inside some pages are gone.
On Processes, a running job's Duration now counts up every second, and Refresh also reloads an open job's detail and the jobs in expanded workflow rows.
Audit History now records only changes someone made, labels each created or deleted record by its kind, marks list items added or removed, and names custom fields and dependencies.
When Audit History cannot be reached, the report says it is unavailable and to try again later, from the first attempt, instead of showing a server error.
Decide who builds, who adds jobs and who changes status. Building a workflow and adding a job now have their own Create permissions, and every job and workflow status change is Execute.
Only the routing a job type can use. The job editor now offers SMAFT and RPA jobs a specific agent only, and MFT, RPA and EASE jobs skip agents whose relay cannot run them.
Keep jobs running when an agent goes down. A job's backup agents now take over: when its primary agent is unavailable, the job runs on the first backup that can take it.
See when an agent starts reaching a different machine. A legacy agent now shows a warning when the host name it reports changes, until you accept the new one.
Jobs running on legacy machines carry through a relay failover. The standby that takes over picks them up and reports their real result; a job it cannot confirm shows as Lost in Failover.
Run a Windows program at the priority it needs. A Windows Command job now has a Process Priority — Normal, High, Idle or Real-Time — for the program and its prerun.
Keep your Episys Run JobFile failure criteria. A UNIX - Episys: Run JobFile job now takes exit criteria, so the RSJ exit codes you treat as success finish OK, as they did in Classic.
Failed Jobs you can trust at a glance. The card now counts what the Failed tile counts, shows each failure's end time and schedule date, and pages and sorts per environment.
Watch a whole workflow's jobs from one notification group. A workflow group can now turn on any job trigger, such as Job Failed, and notifies for every job in its member workflows.
Open a row and it stays put. Processes pauses its automatic refresh while a row is open, so new jobs starting can no longer close the row you are reading.
Spot workflows with failed jobs at a glance. Workflow View shows a red failed-job count beside each affected workflow's status, and a finished one reads Completed with errors instead of Completed.
Report on one day, or everything before or after it. The Job History and Event Details reports' Date filter now offers On, Before and After as well as Between.
Spot every added and deleted record at a glance. Audit History now shows a create or delete as one row, Record Added: <name> or Record Deleted: <name>, and you can filter for either.
See who changed what, and when. The Audit History report is back in Reports, one row per changed value with the before and after, who changed it and where. Run it or download it.
Run an Annual Plan on your Master Holiday Calendar. Pick Master Holiday Calendar (system) as the plan calendar, and the job runs on your organization's holidays, following every change to that list.
Preview a frequency against your real week. The forecast in the frequency dialog now lets you pick the working days and the Master Holiday Calendar, and explains what a non-working day does.
Find any Vision card fast. Vision Config lists every card with its group path, so you can search, sort and page through them, and drag a card's frequencies into priority order.
See the whole Vision board at a glance. Cards show status, environment, description and times on their face, groups show as sections, and card setup now lives in Vision Config under Design.
See every time on the scheduler's clock. Times across Continuum now show in your environment's scheduler time zone, and the job editor names that zone beside the start offsets.
Workflows, scripts, Self Service buttons, properties, resources and thresholds show their workspace's name, not an ID, to people who cannot view the Workspaces tab.
Renaming a threshold, resource, tag or calendar now shows the new name in the workflows and frequencies that use it, at build and in exports, without saving them again.
Imported jobs now point at this environment's scripts, runners, agent pools, agents, connections and cross-workflow dependencies by name, with a warning for anything that matches nothing here.
A $PROPERTY:ADD whose property name holds an unresolved token is now refused and shown as failed, instead of creating a property named with the token's text.
A deployment can no longer switch an embedded sub-workflow to a version with Sub-Workflow off, and a container job you add or change must run a sub-workflow.
A container job now links to and completes with a nested workflow that finished while the container's workflow was not loaded, or that loaded later, instead of staying running.
Jobs no longer fail with Initialization Error after the platform restarts while agent information is briefly unavailable; they dispatch once it is back.
A job aimed at an agent whose relay is lost now waits in Wait Machine, and Kill or a mark clears a job no relay has picked up.
After a Frequencies report download, the note under the button now says when the file leaves out a different number of frequencies than the page lists.
Job History now runs when agent pool names cannot be loaded, showing each job's saved pool name, and its downloads are sorted the same way as the results page.
The Jobs, Calendars, Properties, Resources, Scripts, Service Requests and Thresholds reports now say their list is unavailable and to try again when it cannot be read.
An agent status, job or workflow change made by an event you send is now recorded under your name in Audit History, the agent's Set by and the run's history.
[[SSI.…]] and [[SJI.…]] tokens in a sub-workflow's job parameters now resolve, so migrated jobs that use them run instead of ending in Initialization Error.
[[JI.$EST RUN TIME]] now resolves on every built job, from its frequency's estimated run time, in job parameters, expression dependencies, events and notifications.
Saving a Vision card now confirms each frequency it monitors exists, so a card cannot be saved with a frequency that does not exist or has been deleted.
A new job saves with the default values its required fields show, such as an MFT transfer's Destination File Filter, instead of being refused as missing them.
Re-saving a deployment that already renames a frequency, to add or edit a rename, now saves instead of reporting that another rule already renames onto that name.
A job waiting to retry keeps its planned start when its workflow is released, and a reopened run moves every job's planned start even after a platform restart.
A multi-instance job waiting on a group its predecessor does not have, or on a date before the workflow turned multi-instance, now runs instead of waiting indefinitely.
The deploy dialog's Version field is now a labelled list, rule errors wait until you edit a field or select Deploy, and workflows without container jobs skip the sub-workflow check.
An Advanced transformation rule whose filter quotes a value the rule cannot match is now refused when you save, naming the problem, instead of matching nothing at deploy.
Rename a role without disturbing its members. Members keep the role and its permissions through a rename, nobody is signed out, and the old name keeps working for an hour.
Legacy job results and events are no longer lost when a relay or its connector stops. Results and events from legacy machines are now kept on disk until the platform has them.
Fail over a relay in seconds, without a single alert. A standby relay now takes over in place within about 40 seconds of a crash, and agents stay online through the handover.
Job numbers stay unique through a relay failover. The platform now hands each relay its job numbers, so a standby that takes over never reuses one a running legacy job still holds.
A relay started twice no longer runs twice. A second copy waits as a standby and takes over if the first stops, and the Relays tab shows how many copies are running and where.
See an MCP job's status and statistics while it runs. The Summary tab shows the agent's latest status text, and Agent Feedback holds the task's mix number, process name and times.
Hold z/OS jobs for console messages, tasks, tape and REXX checks. Message, Job/Task, Tape and REXX Pre-run sections join Dataset Pre-run on every z/OS job that submits work.
Read MCP and OS 2200 job output in Continuum. The Output tab lists and opens an MCP job's output, even while it runs, and an OS 2200 job's output once its agent reports it.
Bring z/OS jobs submitted outside OpCon onto your schedule. z/OS Tracked and Queued jobs let your z/OS LSAM report a job it saw start, or is holding, so it runs and finishes in its workflow.
Hold OS 2200 jobs until a prerun or files are ready. Give an OS 2200 Job a prerun runstream, or up to eight file dependencies, that the agent checks before it starts the job.
Run a prerun before any MCP job. Every MCP job type now has a Prerun section: the agent runs a second MCP job first, and the job waits and tries again until it succeeds.
View and edit z/OS JCL without leaving Continuum. Open a job's JCL member from the job editor, the daily job or the diagram, fix it, and save it under your own z/OS user ID.
Catch OS 2200 job mistakes when you save. The OS 2200 Job editor groups its fields in familiar sections, starts a new job with a completion-status rule, and refuses values the agent would reject.
Copy, rename and remove MCP files from a workflow. The new MCP Copy, MCP Change and MCP Remove job types act on a file or a whole directory on your MCP LSAM.
Restart a failed z/OS job from the step that failed. Pick a start step and an end step in the Restart dialog, and set how datasets and GDGs are handled on a restart.
Run started tasks, REXX execs and console commands on z/OS. Three new job types start a CICS or IMS region, run a REXX exec or issue an operator command on your z/OS LSAM.
Read z/OS SYSOUT as the mainframe lays it out. The Output tab lists a z/OS job's SYSOUT in the agent's order, keeps its spacing, and says plainly why output cannot be listed.
Start your OS 2200 runstreams from Continuum. The OS 2200 Job type starts an ECL runstream on your OS 2200 LSAM, with your run card values, tokens and completion-status rules.
See where each run was sent and what it sets off. Job History now shows each job's agent pool and the events it raises, with the condition for each, and filters on both.
React to a z/OS step abend while the job is still running. LSAM Feedback events fire on every z/OS job type's step completions and messages, and the editor warns when * or ? won't match.
Deleting or changing a custom field no longer leaves workflows that cannot be saved: a save checks only the values it changes, and removes values for deleted fields.
A role's permissions and description can now be changed while your tenant's first role sync runs; creating, renaming or deleting a role is still refused until it finishes.
Stopping or restarting a relay cleanly now closes its connections to OS 2200 and BIS machines first, and reconnects to them back off instead of repeating at once.
Legacy jobs running when a relay host loses power or crashes are now picked up again when the relay comes back, instead of never reporting their result.
Turning off Sub-Workflow on a workflow that a container job uses is refused, in the editor, the API and import, so no container is left pointing at a regular workflow.
A job with a blank Maximum Runs and a latest runtime now keeps rerunning until that time, as its editor shows, instead of stopping at its next success.
Deleting several scripts, environments or custom fields now deletes only the rows you can see, and names every one that could not be deleted, with the reason.
Deleting an environment now releases the resource units it held, and Current Allocations lists inactive and deleted environments so you can release their units.
A Vision card action holding a Card token that is not written correctly, such as [[CI.]], is now refused when you save the card, instead of never running.
The job editor and workflow save now refuse property names that differ only in case, naming both, instead of letting the job fail when it is later edited.
Every frequency on a deployed workflow's schedule now builds its jobs, including frequencies an imported workflow names only by name or by an upper-case ID.
Use All resource dependencies shared across environments now take and give back the resource reliably, including one that names its resource only by name.
A recurring container whose nested run was being reset when the platform restarted now finishes the reset and recurs, instead of having its nested run cancelled.
Custom field defaults reach your jobs. A job with no value gets the field's default when its workflow is saved or imported, so built jobs carry it and required fields with one are satisfied.
Resize table columns, and they stay that way. Drag a column header's edge to widen or narrow it on most lists; your browser keeps the width, and a double-click puts it back.
Upgrade a relay by re-running its installer. A re-run keeps the relay's existing settings, and on Windows the installer stops the relay, swaps in the new files and starts it again.
See which build every relay is running. The Relays tab's Version column shows each relay's build as it last checked in, and sorts in release order, so a relay that needs upgrading stands out.
Run your Unisys MCP programs and WFL jobs. MCP Run and MCP Start run a program or start a WFL job on your MCP LSAM, under an MCP usercode, with the fail codes you set.
Run your z/OS batch jobs from a workflow. The z/OS Batch Job type submits a JCL member through JES on your z/OS LSAM, with a condition-code range and Step Control rows.
Find and fix stale SAP job numbers.Check SAP Job Numbers, in the workflow toolbar, compares every SAP R/3 job's stored number with SAP and saves the corrections you choose as new versions.
Pick SAP jobs, servers and process chains from SAP. Select Search… beside an SAP job's fields to look up a job and its number, a server or a process chain on the job's primary agent.
Restart only the SAP child jobs that failed. Pick the failed children of an SAP R/3 job on its page and restart just those, without running the parent or finished children again.
Run your SAP BW process chains from a workflow. The new SAP BW Process Chain job type starts an existing chain by its technical name and records the agent's result code.
See why an SAP job failed, in plain words. A failed SAP R/3 or SAP BW job's Agent Message now says what its exit code means, followed by the agent's own detail.
Start your SAP background jobs from a workflow. The new SAP R/3 Job type starts an existing SAP R/3 or CRM job by name and number, and records the agent's result code.
See when a legacy job is still in its prerun. A legacy LSAM job now shows Pre-run Active while its agent runs the prerun, and Late to Finish and Max Run Time still apply.
Read a legacy job's output while it is still running. The Output tab lists a running legacy job's files instead of waiting for it to end. SAP R/3 jobs excepted.
Move self service buttons between categories by dragging. Organize Self-Service Buttons now refiles a button with a drag, adds new categories in place, and shows each category on a solid colour band.
React to what a legacy agent reports while the job runs. An LSAM Feedback event fires when an SAP child job or a status message matches your pattern, even mid-run.
Rename a workflow without a new version. Select Rename in the workflow's settings: the new name is saved at once, and the version history stays as it was.
A workflow save or import now refuses a custom-field value that does not fit its field, naming the job, field and value, instead of silently replacing or dropping it.
A job the platform has started now always moves on to Running, rather than sometimes staying in Wait to Start, where it could be started a second time.
The Frequencies report now says which workflows it left out and why, and keeps a workflow's other frequencies when its Annual Plan calendar was deleted.
A script's description can now be cleared, and a rename or restore from an out-of-date copy is refused as a conflict instead of overwriting a colleague's change.
A workflow save now checks that you can view each threshold or resource whose dependency it adds, changes or removes, while the rest of the workflow stays editable.
Roles now govern running work too. Acting on running workflows and jobs, reading the Event Log, Vision's live view, and changing environments, saved searches or saved reports all now check the person's roles.
Grant each permission exactly where it applies. A role now holds its own permissions per workspace and per environment, so one role can edit workflows in one workspace and only view them in another.
Roles, event identities and workspaces now share one Permissions page. Each has its own tab under Administration → Permissions, and the Workspaces tab shows how many roles and objects use each workspace.
Agents, pools and relays, all in one place. One page now lists every agent in a single table, shows the jobs each is running, and gives pools and relays their own tabs.
Take a machine out of service from a script.$MACHINE:STATUS now marks an agent Marked Offline or Marked Draining, or clears the mark, and $TOKEN:ADD, SET and DELETE apply as $PROPERTY events.
Run your MySQL scripts from a workflow. The new MySQL job type runs a script file with the mysql client on a SQL LSAM machine, with user variables and exit criteria.
Run Oracle, ODBC and OLE DB work from a workflow. New Oracle and Other DB job types run SQL*Plus scripts, and SQL through ODBC or OLE DB, on a SQL LSAM machine, with exit criteria.
Wait for a file on IBM i. The new IBM i - File Arrival job type watches for a DB2 file or IFS path, with a creation window, a stable-size wait and authority checks.
More of your IBM i work runs from Continuum. New Operator Replay, Restricted Mode and FTP job types, plus message monitors, spool-file actions and dynamic variables on IBM i jobs.
Run your SSIS packages from a workflow. The new MS SQL DTExec job type runs a package from msdb, the file system or the SSIS catalog on a SQL LSAM machine, with exit criteria.
Schedule by the quarter, and pick several days at once. Run a job on a weekday or the first, middle or last day of each quarter, or on several weekdays or days of the month.
Keep one holiday list for every workflow. Enter your holidays once in the Master Holiday Calendar, and every workflow set to Use Master Holiday Calendar treats them as non-working days in forecasts and builds.
A newly registered relay or Universal Agent now shows Offline until its software is installed and checks in, instead of showing Online before anything is running.
A pool or legacy group that still has agents can no longer be deleted: its Delete is disabled, and a request that gets through is refused with the agent count.
Windows File Arrival jobs whose file name is wrapped in double quotes, as migrated jobs often are, now run on a current relay instead of failing at dispatch.
IBM i jobs now receive quotes, ampersands and angle brackets exactly as typed, so a quoted value in a Call or an FTP Remote System reaches the machine intact.
An IBM i job's Termination now shows the message ID its agent reported, such as SMA0036 or SMA0037, and an Exit Description event can match it with Contains.
A job assigned to a specific agent that a transform rule changes now runs on the rule's target agent, including a legacy agent behind a different relay.
Imported workflows and frequencies now point at this environment's calendars, frequencies, thresholds, resources and tags, so they deploy with their frequencies and show as referenced.
After a page error, the rest of Continuum keeps working, the tabs in Agents, Scripts, Toolkit, Permissions and Vision stay usable, and Retry reloads a page that failed to load.
After a page hits an error, opening another page in the same area now shows it, and the message says the page hit an error.
Rebuilding a workflow instance with overwrite now stops its old nested workflow runs as well as the instance itself, so only the replacement starts jobs.
A job inside a container's nested workflow with a same-day dependency on another workflow now waits on that workflow's first instance, as it would in a top-level run.
A job whose name is longer than 40 characters can now be added to a running workflow instance and edited there, up to the same 255 characters Design allows.
An ad-hoc instance added by $JOB:ADD to a job with several instances now takes the properties it leaves out from the instance whose first property matches.
A late deadline set as a fixed time through the API now moves with each recurrence or retry, so later runs are not reported late against the first run's time.
A script or workflow opened while your browser is offline now shows a loading placeholder until its content arrives, instead of a blank form you can type into.
Self Service Configuration's Active column and a button's Deploy modal now show only environments your role may view deployments in; Active shows No access if your role lacks that permission.
The forecast in the Frequencies dialog now comes from the same calculation as the build, so it applies the non-working-day action, exclusion calendars and working-day offsets.
A workflow's reference to a deleted calendar, frequency, threshold, resource or tag no longer moves to a new object of the same name; a required one refuses the save.
Saving a workflow now checks its frequency, threshold, resource and tag references, so Cross Reference lists the workflow and the objects it uses cannot be deleted from under it.
The workflow editor now flags a browser draft that a newer version has overtaken, and refuses to commit it over that version; Discard changes loads the latest.
In a workflow editor window narrower than 1,280 pixels, the toolbar's icon actions move into a More actions menu, so every action stays visible and selectable.
Renaming a workflow, saved or not, keeps the dependencies between its own jobs drawn and resolved as its own, including dependencies on runs built under the old name.
Undo and Redo on the workflow canvas bring back a job with its job type and every setting, and a dependency with its frequency, so the workflow saves afterwards.
Connections are protected by a role now. Each action on a connection needs the matching Connections permission, granted on a role scoped to all workspaces.
Legacy agents can no longer be registered with nowhere to send their events. The default event environment is required and checked when you enter it, and one action fixes every agent still missing one.
See which agent ran a pool job. A completed least-tasked pool job now names the agent that claimed it, instead of only naming the pool it was queued on.
Accented characters now reach your legacy agents intact. The legacy connection uses a single-byte code page, 1252 unless you set another, so command lines, script bodies and file paths keep them.
Run your existing SQL Server Agent jobs from a workflow. The new MS SQL Job type starts a job on a SQL LSAM machine and reports its outcome, or just monitors one already running.
See what each attempt at an event used. An event's detail now has an Attempts section listing each processing attempt and the property-token values it applied, with values you cannot view masked.
Read the event log by name, not by code. Each event shows its display name over the raw $TYPE, and the Event Type filter lists names in category order, searchable by type or category.
Four places that define an event now share one form. Job events, workflow completion events, Self Service buttons and notification actions: pick the event type, fill in its fields, and it is checked on save.
A SEQ job's sequence number lands in a property later jobs can read. It is written to the global property SEQ-<identifier> when the agent has local event credentials and a Default Event Environment.
See a job's past runs.Job Execution History, on a job's Status History tab, lists the job's runs across up to 366 days, with each run's details and notes anyone can add.
See the agent's own reason when a legacy job fails. When a job's Termination shows its exit code, an Agent Message row on its Summary tab shows the failure text the agent sent.
A job can now claim a whole resource. Select Use all on a resource dependency and the job waits until every unit is free, holding the resource against jobs of the same or lower priority.
A job's Priority now decides which ready jobs the platform takes first. Set Priority on the job's frequency, from 0 to 32767; higher-priority jobs are dispatched and checked for their waits first.
Scope an event to one of the job's frequencies. The Fire Only For Frequency (Optional) picker offers the job's frequencies, and an event scoped to a frequency you removed is flagged and refused on save.
A relay now connects to every legacy agent assigned to it, not just the first 50, and keeps track of its running jobs if it stops while saving its state.
A legacy job now sends its real job, workflow and frequency names to the agent, so the agent's job-name variables carry the job's name instead of an internal identifier.
An event's error now shows the property token you wrote, not its value, when another service's refusal quotes the value changed in case, URL-encoded or as an ISO date.
Editing a finished job's events no longer re-sends the ones already delivered, so an added notification no longer brings duplicates of the rest with it.
The Add Job to Workflow Instance dialog now shows Will Replace for each finished instance of a multi-instance job, including one that runs on each agent of a group.
A secret copied into an existing property by a $PROPERTY:SET or $PROPERTY:ADD event now stays encrypted, even if that property's encryption is turned off at the same moment.
Dependencies and conflicts now follow a renamed workflow by identity, both across workflows and between its own jobs, and stay on it when another workflow takes the old name.
An event command whose field holds a property token containing a comma now fires against the right fields, instead of treating the token's comma as a field separator.
A Relative start or latest-start offset is now measured from the workflow's actual start, including after the workflow is released or started from hold.