[object Object]

Freshservice gives you a Problem object. Most teams create them, never link them to incidents or changes, and wonder why nothing improves. Problem management only works as a loop.

Every problem record must have:

  1. Originating incident(s) attached
  2. Root cause statement (not a symptom)
  3. A linked change request implementing the fix
  4. Known error article published, even if the fix is pending

Skip any of these and the problem is bookkeeping, not engineering.

Use the Associate Incidents action aggressively

When ten incidents show the same symptom, do not solve each. Open a problem, associate the incidents, and announce the workaround in the linked Known Error article. Future incidents matching the symptom auto-link via dispatcher rule.

Dispatcher rule: Incident subject contains "VPN drops every 30 min"
  → Tag: KE-2026-014
  → Link to Problem: PRB-1142
  → Send canned reply: workaround steps

Root cause must be a sentence, not a category

“Network” is not a root cause. “DHCP lease time of 30 minutes on subnet 10.4.0.0/16 conflicts with VPN keepalive of 35 minutes” is. Configure the Problem form with a required root_cause_statement text area, minimum 80 characters.

Tie the change request explicitly

Link a Change record from the Problem before you mark it Resolved. Use the API if your team forgets manually:

PUT /api/v2/problems/{id}
{ "associated_change_id": 4471 }

Close the loop with metrics

Track “incidents reopened post-problem-resolution.” If above 5%, your root cause was wrong, not your fix. Re-open the problem.

Pitfalls to avoid

  • Creating problems from a single incident (premature)
  • Closing the problem when the change is approved, not implemented
  • Letting Known Errors expire without review

What to do this week

Audit every open problem for the four required links, add the dispatcher rule for top recurring symptoms, and add a required root cause field with character minimum.

[object Object]
Share