chore: amend ADR 0092 and retire webModeler.persistence.deploymentStrategy after ephemeral-volume switch

[eamonnmoloney]

Context

The per-version PRs #6027 (8.9), #6406 (8.10), and #6408 (8.8) switch the Web Modeler restapi /tmp volume from a shared chart-managed ReadWriteOnce PVC to a per-pod generic ephemeral volume. The /tmp mount is disposable pod-local scratch, so a per-pod volume is the correct primitive and it removes the Multi-Attach upgrade deadlock at the root — without needing the deploymentStrategy: Recreate opt-in introduced by #6014.

This supersedes ADR 0092 for Web Modeler, but the ADR and the value knob are still in place. crev flagged this on all three PRs (P1 on #6027).

Follow-up work (human-owned)

• Amend ADR 0092 (human-authored, per repo rules): remove Web Modeler restapi from the chart-managed-RWO-PVC components table for 8.8/8.9/8.10; update Applicability by version; narrow the remaining Recreate rationale to the existingClaim-with-RWO case; reference fix(8.9): back Web Modeler restapi /tmp with a per-pod ephemeral volume #6027/fix(8.10): back Web Modeler restapi /tmp with a per-pod ephemeral volume #6406/fix(8.8): back Web Modeler restapi /tmp with a per-pod ephemeral volume #6408 as the implementing changes. (8.7 is out of scope — its /tmp is a hardcoded emptyDir, no persistence option.)
• Decide whether to retire webModeler.persistence.deploymentStrategy. The PRs already narrowed the guard/@param/schema text to the existingClaim case; full removal of the knob is deferred here. If kept, it only matters when existingClaim points to an RWO volume.
• Apply the same ephemeral-volume pattern to the other RWO-PVC components: Connectors (fix(connectors): expose opt-in Recreate deployment strategy for RWO persistence #6268), Identity (fix(identity): expose opt-in Recreate deployment strategy for RWO persistence #6269), Optimize (fix(optimize): migrate hardcoded Recreate to opt-in deployment strategy pattern #6270) — all back /tmp with the same shared-RWO anti-pattern across 8.8/8.9/8.10.

References

• PRs: fix(8.9): back Web Modeler restapi /tmp with a per-pod ephemeral volume #6027, fix(8.10): back Web Modeler restapi /tmp with a per-pod ephemeral volume #6406, fix(8.8): back Web Modeler restapi /tmp with a per-pod ephemeral volume #6408
• Supersedes mechanism from: fix(8.8,8.9,8.10): allow Recreate strategy for Web Modeler persistence #6014
• Issue: [ISSUE] create a scenario values.yaml file and a guide for enabling persistence #4767, SUPPORT-30069

[distro-ci]

🧭 Triage Assessment

Field	Value
Work type	TASK
Kind	chore
Severity	⚠️ Mid
Likelihood	🔶 Mid
Confidence	0.91

📝 Rationale

This issue is within SM Platform Distribution scope because the team-boundaries document assigns Helm Charts, the camunda/camunda-platform-helm repository, and issues related to Self-Managed Helm Chart configurations to Distribution. It is best classified as TASK rather than BUG because the underlying Web Modeler runtime failure was already addressed by the per-version ephemeral-volume work: the chart-managed /tmp path now uses a per-pod ephemeral volume, the shared single-attach PVC is removed, and the install-Pending / upgrade Multi-Attach failure plus the deploymentStrategy: Recreate workaround are described as superseded, leaving this issue as follow-up maintenance to align ADR 0092 and the remaining values surface with the new behavior. The best kind is chore, not docs, because the non-feature triage guide defines chore as maintenance / cleanup / tech debt work, and this issue is not documentation-only: ADR 0092 still normatively scopes Web Modeler into the opt-in persistence-scoped strategy pattern, while the current Web Modeler template still exposes webModeler.persistence.deploymentStrategy and its 8.10 override path, so the follow-up spans documentation plus configuration-surface cleanup. Severity is mid because this is more than a cosmetic/local cleanup: the task-triage guide maps medium impact to clear productivity, reliability, or workflow improvement, and removing stale ADR/API guidance reduces real maintenance and user-configuration friction, but the issue does not itself describe a live runtime outage, data loss, or security problem. Likelihood is mid because the stale guidance and redundant knob are repeatable for the subset of contributors or users touching Web Modeler persistence or upgrade behavior, which aligns better with an occasionally encountered / scoped problem than a broadly frequent one. I did not find a strict duplicate already tracking this exact cleanup; the closest related issues are #6268, #6269, and #6270, which track the same ADR-0092 / deployment-strategy follow-up pattern for the other PVC-bearing components, plus #4767 as the original Web Modeler persistence umbrella.

🔍 Similar Issues

#6268, #6269, #6270, #4767

---

🤖 This issue was automatically triaged by an AI-based system.
✋ If you believe this assessment is incorrect, update the labels or add the triage:manual label for human review.