Change service hours without silently rewriting live commitments. Decide which requests keep their existing deadline, which need an authorized recalculation and what evidence closes the change.
A team adds Monday to its holiday calendar. New requests receive deadlines that skip the closure. Some older requests still show Monday as their due date. A paused request changes again when it resumes. The calendar edit succeeded, but the service operation now has several different interpretations of the same change.
SLA calendar change control is the process for assessing, authorizing, applying and reconciling a revision to the working-time calendar behind a service target. Its work object is the calendar-change case, linked to every affected request and measurement cycle. It does not replace the request’s delivery status or authorize a change to the underlying service agreement.
Separate the calendar, the clock and the commitment
A calendar defines eligible working windows. A clock applies a target and event rules to a particular cycle. A commitment is the deadline the service owner is accountable for delivering or explicitly revising. Changing one does not prove that the other two changed correctly.
In Dynamics 365 Customer Service, Microsoft’s guidance on schedule changes says editing working hours does not itself update existing KPI instances’ warning and failure times. Later events can cause recalculation; it also warns about mixed-calendar calculations for instances previously paused or resumed. Those are documented behaviors for that system, not a universal rule.
Intercom’s office-hours guidance likewise says that including a formerly excluded day does not automatically update existing SLA targets unless the SLA is reapplied or overridden. Verify the behavior of your actual platform and configuration before deciding how to treat live work.
If the timing rules themselves are unclear, first use the SLA clock specification to define start, pause, resume, stop and counted minutes. This article addresses the next problem: releasing a calendar revision while retaining an evidenced disposition for existing cycles. Moving requests to a different process definition is a separate in-flight workflow migration decision.
Open one change case and establish its boundary
Trigger review when working days, daily windows, holidays, calendar time zone or calendar assignment is proposed to change. A corrected holiday entry still needs impact review if the old entry has already influenced deadlines or reporting.
The change coordinator collects the current and proposed calendar revisions, reason, intended effective instant with time zone, affected services, governing target rules and requested treatment of existing work. Missing scope or an unverified current calendar keeps the case in Impact review; the coordinator owns clarification.
- Calendar owner: verifies the actual working windows, closure dates and intended revision.
- Service owner: authorizes treatment of live commitments within their authority and obtains any additional required agreement.
- System operator: proves the engine’s behavior, executes the approved change and establishes the actual result.
- Receiving work owner: accepts the current deadline, next action and escalation responsibility.
- Change coordinator: maintains coverage, exceptions, acceptance and closure evidence.
Use an authoritative selection event—for example, cycle start before or after the effective instant—not the time an operator happens to open a request. Record how an old request beginning a new cycle will be treated. A new measurement cycle and a newly received request are not necessarily the same population.
Find every dependency, not just the default calendar
Inspect service-to-calendar assignments, team or regional overrides, warning rules, escalations, views, notifications and reports that use business time. Zendesk’s schedule documentation shows that business-hour rules and metrics can depend on the selected schedule. A change review therefore needs more than a screenshot of the holiday list.
Build the affected-cycle register from authoritative records. State the query, statuses included, observation time and any inaccessible population. Capture request and cycle IDs, calendar revision, target, start time, current state, pause/resume history, original due time, original breach result, owner and downstream warnings already sent. An incomplete inventory blocks broad release; a small sample is a test set, not coverage proof.
Choose treatment before applying the revision
The service owner decides which obligation should apply; the operator establishes whether the system can represent it safely. Neither a later due date nor an unchanged badge is evidence of an approved decision.
| Affected population | Treatment to authorize | Evidence required |
|---|---|---|
| Cycles beginning after the approved boundary | Use the new calendar under the approved selection rule. | New cycle points to the intended calendar and produces tested warning and due times. |
| Existing active cycles whose commitments remain valid | Retain the original commitment and supported timing treatment. | Original due time preserved; later events cannot silently substitute another interpretation. |
| Existing active cycles selected for revised treatment | Recalculate only through a supported, tested mechanism with service-owner approval. | Original, approved and observed due times; reason, authority, calculation basis and accepted next owner. |
| Paused or previously resumed cycles | Review separately before any recalculation. | Complete relevant event history and a test proving how old and new calendar windows are used. |
| Completed or already breached cycles | Preserve the original result; route genuine corrections through authorized review. | Before/after results, correction rationale and distinct reporting treatment—not a silently improved score. |
| Unknown calendar assignment or uncertain operation outcome | Hold the change disposition for reconciliation. | Named owner establishes actual assignment/result before another modifying operation. |
If preserving an old calendar is not technically supported, do not claim a prospective-only edit is safe. The service owner and operator must choose a supported alternative or defer the release. Changing display text alone does not control the timer behind it.
Move the change through explicit release states
Use Proposed → Impact review → Approved → Executing → Reconciling → Accepted for the change case. A rejected proposal ends with its reason. A failed gate moves to Held with a named resolution owner and the state to return to after verification. Do not use these labels to pause the underlying service clock.
Approval to execution: hand over a bounded instruction
The coordinator sends the operator an approved packet only after the service owner has accepted the exact revision, selection boundary and cohort treatments. Include calendar snapshots, cycle IDs, original deadlines, expected results, test evidence, permitted operation, stop conditions and exception owners. The operator accepts by confirming access, supported execution and a fresh state check. Missing or conflicting instructions return to the coordinator; they do not become implementation choices.
Execution to operations: accept the actual commitments
The operator sends each receiving owner the request/cycle identity, actual calendar, original and approved due times, observed system due time, warning disposition, remaining work and unresolved exceptions. Acceptance requires that the owner can access the record, understands the next action and confirms the current commitment. A successful administrative response is not this handoff.
Keep any customer-facing revised promise consistent with the authorized decision. Record who communicated it and what was actually communicated. A failed notification remains an owned communication exception, not evidence that the requester accepted a different deadline.
Test the events that can change a result later
Before production release, record expected and observed behavior for representative active, paused, resumed, completed and breached cycles. Include a fresh cycle, a cycle exactly at the boundary, multiple calendar assignments, a closure adjacent to a weekend and any relevant daylight-saving transition.
Most importantly, test after the edit: resume a paused test cycle, change a field that can reapply the target, and complete a cycle. Check whether the engine preserves, replaces or mixes timing inputs. Do not transfer a vendor-specific recalculation technique into another platform without documented support and testing.
- History is missing or calculation disagrees: the operator owns investigation; hold revised treatment until the event history and expected result are verified. Escalate an unresolved interpretation to the service owner.
- A cycle completes during selection: re-read its authoritative state before execution. Reclassify it under the approved completed-cycle rule; retain the observation and decision.
- The operation times out: stay Reconciling. Inspect the original operation and affected records before repeating a write. Never create replacement service requests to escape uncertainty.
- A second calendar edit appears: the calendar owner identifies the controlling revision. Hold conflicting release instructions until one current decision and impact set are approved.
- The receiving owner is unavailable: the service owner names an authorized backup. Resume the handoff only after access and acceptance are confirmed.
- The release must be reversed: stop expansion and obtain an approved recovery plan. Restoring old settings does not prove that recalculated targets, sent messages or historical reports reverted.
Service delivery should continue under its authorized operating instructions while change exceptions are resolved. Where an actual promise is at risk, escalate it immediately; do not hide the risk by pausing measurement or replacing the original deadline.
Worked example: a closure creates two deadline populations
Illustrative design example, not a customer result: an internal review service counts Monday–Friday, 09:00–17:00 in America/Chicago. A cycle starts Friday at 16:00 with a four-working-hour target and no pauses. Under calendar v1, one hour accrues Friday and three Monday, giving an original due time of Monday at 12:00.
The organization proposes a Monday closure in v2. A calculation using that closure across the whole cycle would instead reach four counted hours Tuesday at 12:00. That arithmetic does not authorize changing the existing promise.
The service owner chooses new cycles on v2 and retains this existing cycle’s Monday commitment. Before approving release, the operator must prove that later engine events will not silently move that old cycle to Tuesday. If the system cannot preserve the chosen treatment, the case stays Held for an authorized alternative.
A second existing cycle has pause/resume history. It is not assumed equivalent to the first. The operator reconciles its history and tests the exact behavior separately. It remains unresolved until the service owner approves a supported treatment and the receiving owner accepts the observed result. The change cannot be called fully accepted merely because new requests look correct.
Close against a reconciled population
At release acceptance, every selected cycle must have one evidenced treatment: retained, revised-and-accepted, completed with original result preserved, or explicitly unresolved with an authorized owner and follow-up. Reconcile those counts to the selection total. Keep exceptions visible; do not omit them from a success denominator.
- Demand and capacity: affected cycles by calendar/state versus review and acceptance capacity.
- Flow and queue health: time to accepted disposition, oldest Held item and overdue owner acceptance.
- Quality and reliability: mismatches between approved and observed due times, unknown outcomes and unapproved historical-result changes.
- Outcome: requests delivered against the accepted commitment; show original and revised commitments separately where applicable.
Acceptance means every case has an approved disposition and the actual system behavior is understood—not necessarily that every underlying request is finished. An unresolved discrepancy needs explicit acceptance as an exception, not a fabricated technical pass. Review recurring mismatches to improve the particular selection rule, test or handoff that failed.
Where Workhint fits
The Workhint Workflow Engine describes approval routing by role and requiring details or files before progression. Those are relevant building blocks for calendar-change intake and review. Workhint’s Workflow Automation Software page describes configurable forms, stages, tasks and event-based notifications for coordinating the follow-up.
These pages do not establish native SLA-calendar arithmetic, live-cycle recalculation, immutable snapshots or automatic historical-report correction. Treat those as requirements to verify in the selected timing system. Workhint’s relevance here is coordinating accountable decisions and handoffs, not assuming that a workflow stage changes a deadline correctly.
Build This in Workhint
Use this configuration brief: “Coordinate changes to our service calendar. Outcome: every affected cycle has an authorized and accepted treatment. Define calendar-owner, service-owner, operator, coordinator and receiving-owner roles with limited permissions. Collect revision, reason, effective boundary, affected services, cycle references and original/approved/observed due times. Track Proposed, Impact review, Approved, Executing, Reconciling, Accepted, Held and Rejected separately from service status. Require complete coverage, test evidence and approved cohort treatment before execution. Assign missing-data, discrepancy and communication tasks; notify owners at decision and acceptance events. Retain actor, time, decision, evidence and prior-result history; verify audit behavior. Provide cohort, mismatch, held-age and acceptance views using the measures above. Validate calendar calculations and execution controls separately before enabling any unattended change.”
FAQ
Does adding a holiday automatically change existing deadlines?
Do not assume it does. Establish the documented behavior of the actual system, including what happens on later resume, reapplication and completion events. Then compare observed results with the approved treatment.
Can a revised calendar erase an earlier SLA breach?
It should not silently overwrite the original business result. Preserve that result and route genuine calculation corrections through authorized review with before/after evidence and clear reporting treatment.
Is a change hold the same as pausing a service request?
No. A release hold stops an unverified calendar-change action. A service pause has its own permitted reasons and evidence under the clock specification; uncertainty about a release is not automatically one of them.

Leave a Reply