transactions_skill
- Python
2
GitHub Stars
1
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill pluginagentmarketplace/custom-plugin-sql --skill transactions- SKILL.md962 B
Overview
This skill teaches database transactions with a focus on ACID properties, isolation levels, and practical concurrency control. It covers patterns and tools for building reliable, consistent data operations in single-node and distributed systems. You will learn trade-offs and design choices that affect performance and correctness.
How this skill works
The skill inspects common transaction scenarios and explains how atomicity, consistency, isolation, and durability are achieved in SQL databases. It breaks down isolation levels, locking strategies, MVCC behavior, and failure-handling patterns such as two-phase commit and sagas. Concrete examples and patterns show how to apply savepoints, nested transactions, and deadlock recovery in application code.
When to use it
- Implementing multi-step business operations that must be atomic
- Choosing an appropriate isolation level to balance consistency and throughput
- Designing distributed workflows that span multiple services or databases
- Troubleshooting deadlocks, lock contention, or stale reads
- Optimizing transactional throughput in high-concurrency environments
Best practices
- Start with the weakest isolation that meets correctness, escalate only when necessary
- Keep transactions short and move nonessential work outside the transaction boundary
- Prefer optimistic concurrency for low-conflict workloads and MVCC-capable DBs
- Use savepoints for partial rollbacks and sagas for long-running distributed processes
- Monitor locks and transactions to detect contention and tune indexes or queries
Example use cases
- Bank transfer: ensure atomic debit/credit across accounts with rollback on failure
- Order processing: combine inventory, billing, and shipment updates with compensating actions
- Concurrent counters: choose locking or optimistic updates to avoid lost updates
- Schema migrations: run safe DDL in transactions or use phased deployments
- Microservices: coordinate state changes across services using two-phase commit or saga pattern
FAQ
Use serializable when strict correctness is mandatory and anomalies cannot be tolerated; expect reduced concurrency and potential performance impact.
How do I handle deadlocks in production?
Detect and log deadlock incidents, ensure transactions are short, order resource acquisition consistently, and implement retry logic with backoff.
When are sagas preferable to two-phase commit?
Use sagas for long-running, loosely coupled distributed workflows where blocking coordinator-based protocols would be impractical; implement compensating transactions for failure recovery.
