azure-workflow-execution-models
Skill4 starsChanged 5 months agoMITView on GitHub
Time-bound reference for comparing Camunda, Azure Functions, Dapr, and KEDA as workflow execution models in Azure-native environments. Use when evaluating execution model candidates for a service orchestration layer, or when assessing claims about these technologies in architecture comparison artifacts. Call for Camunda Python SDK, Zeebe streaming, Dapr CNCF status, Azure Functions timeout, KEDA workload identity, or any freshness check on vendor-specific execution-model facts.
- Installs packages
In this repository101 other files it gives its agents.
- ai-video-understandingSkill
- apple-foundation-modelsSkill
- arch-dddSkill
- arch-decision-recordsSkill
- arch-diagramsSkill
- arch-eventsSkill
- arch-mvp-roadmapSkill
- arch-python-modernSkill
- aws-app-runnerSkill
- aws-cicd-patternsSkill
- aws-security-hardeningSkill
- azure-functions-python-v2Skill
- azure-identity-m2m-authSkill
- azure-service-bus-messagingSkill
- caddy-tls-proxySkill
- compliance-frameworksSkill
- design-system-workflowSkill
- dev-debugging-strategiesSkill
- dev-standardsSkill
- docker-python-poetrySkill
- docker-pythonSkill
- fastapi-patternsSkill
- ffmpeg-laravel-videoSkill
- formik-yup-react-nativeSkill
- frontend-accessibilitySkill
- frontend-angular-aiSkill
- frontend-angularSkill
- frontend-angular-toolingSkill
- frontend-design-systemsSkill
- frontend-material-chatSkill
- frontend-material-design-3Skill
- frontend-svelteSkill
- gcp-cicd-patternsSkill
- gcp-cloud-runSkill
- gcp-cloudsql-infrastructureSkill
- gcp-finopsSkill
- gcp-observabilitySkill
- gcp-security-hardeningSkill
- github-actions-cicdSkill
- healthkit-biofeedbackSkill
- infra-terraformSkill
- ios-app-qualitySkill
- ios-audio-hapticsSkill
- jest-nestjs-testingSkill
- langchain-azure-openai-patternsSkill
- laravel-auth-hardeningSkill
- laravel-ci-mysqlSkill
- laravel-mix-webpackSkill
- laravel-modern-patternsSkill
- laravel-pint-patternsSkill
- laravel-s3-storageSkill
- laravel-sail-dockerSkill
- mobile-app-deploymentSkill
- mobile-e2e-testingSkill
- mobile-offline-syncSkill
- mobile-push-notificationsSkill
- mobile-video-uploadSkill
- mongodb-asyncmongoclient-patternsSkill
- mongodb-atlas-patternsSkill
- nestjs-patternsSkill
- The first 60 shown.
The file
Files
Azure Workflow Execution Models
Research verified as of 2026-05-13. Re-verify before reuse in any client-facing artifact.
This skill is time-bound by design. The "Before You Compare" section is the first stop — corrections here have already invalidated claims in production evaluation memos.
Before You Compare
Seven material corrections relative to common 2024-era evaluation memos. Verify each against vendor docs before citing:
- Camunda official Python SDK exists —
camunda-python-sdk(REST-based, sync/async, OAuth/Basic auth).pyzeebe(community, gRPC) is no longer the canonical Python SDK. - Zeebe gRPC streaming since Camunda 8.3 (Sept 2023) — opt-in; Camunda benchmarks show ~50% p50 / ~75% p99 latency improvement. Poll-only latency-penalty claims that omit the streaming caveat misrepresent Camunda 8.3+ performance.
- Camunda Self-Managed commercial license since 8.6 (Oct 2024) — Zeebe was Apache 2.0 free for production pre-8.6; now requires a commercial license for production deployment.
- Camunda Web Modeler git sync since 8.6 — two-way sync with GitHub, GitLab, Azure DevOps. BPMN-not-in-code risk is now a configuration choice, not structural.
- Dapr graduated CNCF November 2024 — same tier as Kubernetes, Prometheus. "Incubating" claims are factually wrong.
- Azure Functions Premium plan timeout is configurable to unbounded — 30 min is the default, not a hard cap.
- Azure Functions Service Bus extension v4.x retired March 2025 — use v5.x.
Quick Reference
| Candidate | Role | Latest stable | Python story | Key recent change | Churn rate |
| Camunda | BPMN workflow engine | 8.8 (Oct 2025); 8.10 forthcoming | Official REST SDK + legacy pyzeebe (community) | Self-Managed commercial license since 8.6; gRPC streaming since 8.3 | High |
| Azure Functions | Serverless compute | Runtime v4; Python 3.10-3.13 | First-class (isolated-process model) | Service Bus v4.x retired Mar 2025; Premium timeout configurable to unbounded | Medium |
| Dapr | Sidecar runtime for distributed apps | v1.16 (Sep 2025) | Sidecar-agnostic; SDK available | CNCF graduated Nov 2024; OTel-native (portability tax if APM SDK is proprietary) | Medium |
| KEDA | Event-driven autoscaler | 2.19 (stable) | N/A — infra layer | Pod-identity removed in 2.15; migrate to workload identity | Low |
Azure Functions — Additional Verification Items
Two commonly mis-stated facts to pin before citing Functions in feasibility or handoff docs:
- Handler-load isolation check — cheapest CI gate for "is this Function App well-formed" without deploying the host. The one-liner that imports the app and asserts trigger registration lives in
reference.md(and the build-side detail is owned byazure-functions-python-v2). - Dynatrace Functions extension is App-level, not container-level — installed as a
siteExtensionsresource on the Function App (portal or ARM/Bicep), not viaRUN apt-get installin a custom container Dockerfile. Custom containers are compatible with the extension. The common misread is that a custom container image precludes the extension.
Freshness Discipline
Vendor-specific facts age at different rates. Licensing models, SDK ecosystem shifts (official vs community), CNCF tier transitions, broker transport modes, and default-on flags churn annually to sub-annually. Language runtimes and mature CNCF post-graduation contracts are stable for multiple years. When one claim is flagged outdated, sweep all comparable claims — outdated facts cluster. See reference.md for the full fast/slow taxonomy and arch-decision-records for the cluster-sweep rule.
Anti-Patterns
| Anti-Pattern | Pattern |
| Using poll-only latency as a Camunda differentiator without streaming caveat | Qualify: polling is default; gRPC streaming (opt-in since 8.3) reaches push-broker parity |
| Citing Dapr as "CNCF incubating" | Dapr graduated CNCF top-level project November 2024 |
Citing pyzeebe as the canonical Python SDK for Camunda | The official camunda-python-sdk (REST-based) is now first-party; pyzeebe is community |
| Treating Azure Functions 30-min timeout as a hard cap | 30 min is the default; host.json configures it to unbounded |
| Reusing this skill beyond ~12 months without re-verifying vendor docs | Re-run a vendor-doc-first research sweep; high-churn vendors (Camunda, Dapr) shift annually |
Cross-References
azure-functions-python-v2— implementation-side companion for Functions:function_app.pyshape,host.json, startup-state pattern, container deployment, Dynatrace extension wiringarch-decision-records— fast/slow churn taxonomy; cluster-sweep rule; vendor-doc-first sourcingazure-service-bus-messaging— Service Bus trigger details, v5.x extension, session handlingazure-identity-m2m-auth— workload identity for KEDA, Dapr component auth
See reference.md for per-claim source citations organized by technology.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

