Error Mitigation Module (`hierachain/error_mitigation/*`)

Overview

The Error Mitigation module acts as HieraChain's "safety net," ensuring the system always maintains integrity and resilience against hardware, software, or network incidents. This system combines durable journaling techniques, intelligent error classification, and automated recovery scenarios.

Multi-layered Defense Architecture

Validation Layer

Files: validator.py, data_validator.py
- Validator: Checks Block/Event structure according to Ledger guidelines.
- DataValidator: Deep validation of business logic, Arrow schema, and input data validity.
Durable Journaling

File: journal.py
- Uses Apache Arrow for high-speed transaction journaling.
- Append-only mechanism ensures data is never overwritten.
- Guarantees persistence before events are committed to the chain.
Rollback & Snapshots

File: rollback_manager.py
- Manages restore points (Snapshots) for the entire system or individual components.
- Automatic periodic snapshot creation (Auto-snapshot).
- Validates data integrity before performing rollback.
Adaptive Recovery

File: recovery_engine.py
- Automatic handling of network errors and partition detection.
- Consensus state recovery when a Leader fails.
- Integrates AutoScaler to scale resources based on system load.

Error Classification Strategy

The ErrorClassifier not only logs errors but also proposes mitigation strategies based on severity:

Level	Meaning	Suggested Action
INFO / WARNING	Information or minor error	Log & Continue
ERROR	Transaction processing error	RETRY / REJECT
CRITICAL	Data / consistency error	ROLLBACK & QUARANTINE
FATAL	Critical system error	EMERGENCY SHUTDOWN

Transaction Journal

HieraChain uses Apache Arrow IPC format for Journaling to achieve optimal performance:

Durable Write: Events are written to disk and fsync is called before proceeding.
Schema Enforcement: Ensures every journal record conforms to the core event schema.
Replay Ability: When the system restarts after a failure, the Journal can replay uncommitted events to restore state.

from hierachain.error_mitigation.journal import TransactionJournal

# Initialize a secure journal (Path Traversal protected)
journal = TransactionJournal(storage_dir="data/journal")

# Write a durable event
journal.log_event(event_dict)

Recovery Workflow

When an error is detected, the system follows this processing flow:

graph TD
    A[Incident occurs] --> B{ErrorClassifier}
    B -->|Low Level| C[Log & Continue]
    B -->|Medium Level| D[Auto Retry / Recovery Engine]
    B -->|High Level| E[Rollback to nearest Snapshot]

    D --> D1[Network Recovery]
    D --> D2[Consensus Recovery]
    D --> D3[Auto Scaling]

    E --> F[Validate integrity after Rollback]
    F --> G[Replay Journal to restore missing data]

Snapshot Management

RollbackManager supports different Snapshot types:

CONFIGURATION: Backup only system configuration files.
CHAIN_STATE: Backup Main Chain and Sub-Chains state.
CONSENSUS_STATE: Stores View Number and current Leader information.
FULL_SYSTEM: Snapshot of the entire system state.

Error Mitigation Module (hierachain/error_mitigation/*)