⚡ Workflows
Retry & Rollback (resilient task execution)
Ensure tasks complete reliably by retrying on transient failures, rolling back partial changes on permanent failures, and escalating when recovery is impossible.
Bundle files
📄ORCHESTRATION.md— Orchestration
1# ORCHESTRATION.md
2
3## Diagram
4
5 ┌──────────────────────────────────────────┐
6 │ Task Assigned │
7 └────────────────────┬─────────────────────┘
8 ▼
9 ┌──────────────────────────────────────────┐
10 │ Executor Agent │
11 │ (attempt task) │
12 └────────────────────┬─────────────────────┘
13 ▼
14 ┌─────────┐
15 │ Success?│
16 └────┬────┘
17 yes │ no
18 ┌──────────┴──────────┐
19 ▼ ▼
20 ┌──────────────┐ ┌───────────────────┐
21 │ Deliver │ │ Retry budget │
22 │ Result │ │ exhausted? │
23 └──────────────┘ └────────┬──────────┘
24 no │ yes
25 ┌──────────┴──────────┐
26 ▼ ▼
27 ┌──────────────┐ ┌──────────────────┐
28 │ Backoff + │ │ Rollback Agent │
29 │ Retry │ │ (compensate + │
30 └──────┬───────┘ │ restore state) │
31 │ └────────┬──────────┘
32 ▼ ▼
33 ┌──────────────┐ ┌──────────────────┐
34 │ Executor │ │ Escalate to │
35 │ (re-attempt)│ │ Human │
36 └──────────────┘ └──────────────────┘
37
38
39## Workflow
40Retry & Rollback (resilient task execution)
41
42## Objective
43Ensure tasks complete reliably by retrying on transient failures, rolling back partial changes on permanent failures, and escalating when recovery is impossible.
44
45## Roles
46- **Executor:** attempts the task and reports success/failure with context.
47- **Retry Controller:** decides whether to retry (transient) or abort (permanent), manages backoff.
48- **Rollback Agent:** reverses partial changes using compensation actions.
49- **Escalation Target:** human or supervisor notified when all recovery is exhausted.
50
51## When to use
52- Tasks that modify state (deployments, data migrations, API calls).
53- Environments with transient failures (network, rate limits, timeouts).
54- Any workflow where partial completion is worse than no completion.
55
56## Protocol
571. Executor attempts the task and checkpoints state before mutations.
582. On failure, classify error:
59 - **Transient** (timeout, 429, 503): retry with exponential backoff.
60 - **Permanent** (4xx, validation, logic error): skip to rollback.
613. Retry up to N attempts (configurable budget).
624. If retries exhausted → trigger Rollback Agent.
635. Rollback Agent executes compensation actions in reverse order.
646. If rollback succeeds → report clean failure with context.
657. If rollback fails → escalate to human immediately.
66
67## Retry policy
68- **Max retries:** 3 (configurable per task)
69- **Backoff:** exponential (1s, 2s, 4s) with jitter
70- **Timeout per attempt:** 30s default
71- **Circuit breaker:** after 3 consecutive failures, pause all retries for cooldown
72
73## Rollback rules
74- Compensation actions must be idempotent.
75- Execute in reverse chronological order.
76- Each rollback step is itself retried once on failure.
77- Full rollback log with timestamps.
78
79## Rules
80- Never retry permanent errors.
81- Always checkpoint before mutation.
82- Rollback must leave system in a known-good state.
83- All attempts and outcomes logged for audit.
84
85## Deliverables
86- Execution attempt log (attempt #, error, classification)
87- Rollback action log (if triggered)
88- Final status: success | rolled-back | escalated
89- Metrics: attempts used, total elapsed time
90
Lines: 90 | Words: 415
Install
Copy bundle to your OpenClaw workspace.
curl -fsSL https://raw.githubusercontent.com/cerealskill/openclaw-agents/main/install.sh | bash -s workflow retry-rollback ENRate this agent
Loading...
Sign in to rate this agent
Includes
- ✓ ORCHESTRATION.md
Info
- Slug
- retry-rollback
- Lines
- 90
- Words
- 415