# Rollback Strategies {#rollback-strategies}


Roll back failed deployments safely.


## Kubernetes Rollback {#kubernetes-rollback}


```bash
# Instant rollback to previous revision
kubectl rollout undo deployment/gauzy-api

# Rollback to specific revision
kubectl rollout history deployment/gauzy-api
kubectl rollout undo deployment/gauzy-api --to-revision=3

# Check rollout status
kubectl rollout status deployment/gauzy-api
```


## Docker Rollback {#docker-rollback}


```bash
# Run previous image tag
docker service update --image ghcr.io/ever-co/gauzy-api:v1.2.3 gauzy-api

# Docker Compose - revert to previous tag
docker compose pull
docker compose up -d
```


## Database Rollback {#database-rollback}


If the release included migrations:

```bash
# TypeORM
yarn typeorm migration:revert

# MikroORM
npx mikro-orm migration:down
```


## Rollback Decision Tree {#rollback-decision-tree}






```mermaid
graph TB
    Issue[Issue Detected] --> Severity{Severity?}
    Severity -->|Critical| Immediate[Immediate Rollback]
    Severity -->|Major| Assess[Assess Impact]
    Severity -->|Minor| Hotfix[Apply Hotfix]
    Assess -->|Wide impact| Immediate
    Assess -->|Limited| Hotfix
    Immediate --> DBCheck{DB Migrations?}
    DBCheck -->|Yes| MigRevert[Revert Migrations]
    DBCheck -->|No| ImageRevert[Revert Image Only]
    MigRevert --> Verify[Verify Health]
    ImageRevert --> Verify
```






## Pre-Rollback Checklist {#pre-rollback-checklist}


- [ ] Confirm the issue is deployment-related
- [ ] Identify the last known good version
- [ ] Check if data migrations need reverting
- [ ] Notify team and stakeholders
- [ ] Have monitoring dashboards open


## Post-Rollback {#post-rollback}


- [ ] Verify service health
- [ ] Check error rates returning to baseline
- [ ] Document the rollback
- [ ] Root cause analysis
- [ ] Fix forward with new release


## Related Pages {#related-pages}


- [Release Management](https://sourcey-gauzy.pages.dev/workflows/release-management/index.md) — releases
- [Blue-Green Deployment](https://sourcey-gauzy.pages.dev/devops/blue-green-deployment/index.md) — zero-downtime
- [Disaster Recovery](https://sourcey-gauzy.pages.dev/devops/disaster-recovery/index.md) — DR planning