- Which side is the service project
- Request type
- Comment visibility
- Customer / Reporter
- Organizations, participants, approvers (opt-in)
- Sanitization
- What is not synced via fields
- Related
The Jira Service Management section appears in the editor when at least one side is a service project. The pair type shows JSM ↔ Jira or JSM ↔ JSM. Everything on this page lives inside that section unless noted.
Which side is the service project
In a pair with a service project, make the service project Project B: for issues from Project A the app creates a customer request in Project B through the service desk (with a request type and a reporter); issues from Project B get a regular issue in Project A.
Request type
| Setting | Default | Notes |
|---|---|---|
| Auto-map request type by name | on | Picks the request type whose name matches the issue type |
| Issue type to request type | empty | Explicit mapping per issue type; takes the place of the name match |
The issue type itself still comes from the issue type mapping, and its When source type is unmapped policy applies first. If no request type can be resolved, the counterpart is not created.
The portal and request types of the target service desk are looked up across their whole lists, not only the first page.
Comment visibility
JSM comments are either public (visible to the customer on the portal) or internal (agents only). Comment visibility decides what comments written into the service project become:
| Mode | Effect |
|---|---|
| Internal (default) | Comments are internal notes |
| Public | Comments are visible on the portal |
| Preserve source | Keeps the visibility of the source comment |
| Label-based | Public when the issue has the Public label |
If comments are public and email or phone masking is off, the editor shows a Portal safety check warning - see Sanitization.
The sync copies all comments of an issue, not only the first 100: it reads up to 50 pages of 100 comments. If the comment list could not be read to the end, the run fails for that issue instead of syncing a partial list; retry it from the failed items list.
Customer / Reporter
When the app creates a customer request, it needs a reporter.
| Setting | Default |
|---|---|
| Customer policy - Bot (use bot accountId), Selected customer, Original reporter (when possible) | Bot |
| Bot accountId, Customer accountId | empty |
| Add original reporter as request participant | off |
- Bot (use bot accountId) - the bot account is the reporter on every synced request. Customer identities never appear on the target side.
- Selected customer - one fixed customer (Customer accountId) is the reporter.
- Original reporter (when possible) - the source reporter is used when possible; otherwise the bot is used.
Account IDs are left out of exported configurations; re-select them after an import. A bot account that does not resolve to a Jira user blocks the save.
Organizations, participants, approvers (opt-in)
| Setting | Default | Notes |
|---|---|---|
| Sync organizations | off | Matched by name; missing organizations are created on the target and attached to its service desk |
| Sync participants | off | Participants must already exist; unknown accounts are skipped and audited |
| Sync approvers | off | Read-only: differences are audited, approvals are never answered by the sync |
Organizations
Organization IDs differ between sites, so organizations attached to the request are matched by name, across the whole organization list of the target. A missing organization is created, attached to the target service desk and to the target request. User membership inside an organization is not synced - that is site-level admin work.
If the target organization list could not be read to the end, the app does not create the organization: the issue fails instead, so the sync never creates duplicate organizations. Retry it from the failed items list - see Troubleshooting.
Request participants
Participants are compared by account ID. Missing participants are added to the target request; participants removed on the source are removed. An account that does not exist on the target is skipped and audited - the rest of the sync continues. Participants are synced independently of the customer policy.
Approvers (read-only)
The sync never answers an approval. Differences between the approvals of the two requests are written to the audit log only - this is an observability feature, not a transaction. If you need an approval decision mirrored, record it in a custom field and sync that field.
SLA snapshot
Ignore SLA (recommended) is on by default. With it off, Mirror SLA remaining time as a snapshot copies the remaining time (milliseconds) from Source SLA field ID into a numeric Target snapshot field ID. The target SLA timer itself is not restored - the snapshot is informational.
Sanitization
Sanitization scrubs PII out of synced content before it lands on the target. There are two layers: built-in options and your custom rules.
Built-in options
Applied before data is mirrored to the portal-visible side. All options are off by default.
| Option | What it does |
|---|---|
| Mask emails | Masks email addresses in mirrored text |
| Mask phone numbers | Masks phone numbers |
| Mask URLs | Masks http://... and https://... links |
| Replace @mentions with @bot | Mentions point to the bot instead of the original user |
| Replace assignee, Replace reporter, Replace other people fields | keep, bot or app each |
| Exclude fields | Fields never mirrored |
These cover most “personal data out of the other project” scenarios. If comments are public and email or phone masking is off, the editor shows a Portal safety check warning.
Custom sanitization rules
For vendor IDs, internal codes, or project-specific personal data you can write Custom regex rules. Up to 32 rules per sync.
| Field | Notes |
|---|---|
| Name | Shows in audit logs; up to 80 characters |
| Pattern | Regex source, up to 512 characters |
| Flags | g, i, m, s, u |
| Replacement | Up to 256 characters; supports $1 backreferences |
| Scope | Field, Comment or All |
| Field allowlist | Optional - limit the rule to specific fields |
| Enabled | Disabled rules are kept but not applied |
Backtracking protection: patterns at risk of catastrophic backtracking (for example nested quantifiers like (a+)+) are rejected on save. At run time, a rule that takes longer than 100 ms on one value is skipped and audited.
Preview on sample: shows the effect of the rules on up to 10 recent issues. Nothing is written. Available after the sync is first saved.
Order of application
Field value resolved → built-in options → custom rules → write to target
Custom rules run after the built-in options. Within custom rules, order is the order in the table - drag (or move up / move down) to re-order.
What is not synced via fields
These JSM-specific objects are synced via dedicated endpoints, not via the regular field sync:
| Object | Path |
|---|---|
| Request participants | /rest/servicedeskapi/request/{key}/participant |
| Organizations | /rest/servicedeskapi/request/{key}/organization |
| Approvals | /rest/servicedeskapi/request/{key}/approval (read-only) |
| SLAs | /rest/api/3/issue/{key}?expand=sla (read snapshot only) |
The Field schema diff tags these as JSM special (synced separately) so you do not accidentally add them as regular field rules.
Related
- Quick Start
- Permissions setup - scopes needed for each JSM extension
- Typical scenarios - scenario B uses the Bot customer policy and custom sanitization rules
- Troubleshooting - JSM request type, JSM customer and JSM scope missing error categories
