SRE Manual

This page introduces observability and operational entry points for Univer services, including metrics, logs, and common diagnostics.

Observability

Univer provides:

• Prometheus metrics
• Service logs (stdout + persistent volumes)

If you enable the built-in observability stack, you can view dashboards in Grafana. Otherwise, export dashboards into your own system.

Metrics

Built-in Grafana dashboards cover:

• Infrastructure: RabbitMQ, Redis, RDS
• Collaboration: collaboration-server
• Core service: universer
• Import/export: exchange

You can use these dashboards to monitor performance, availability, and business KPIs. For external systems, download dashboards here:

https://release-univer.oss-cn-shenzhen.aliyuncs.com/release/univer-grafana-dashboards.tar.gz

Dashboard Samples

1. Golden signals (QPS, SLI, error distribution)

2. Infrastructure performance & availability

3. Collaboration metrics

4. USIP integration metrics

Service Logs

Docker Compose

Logs are written to stdout and persistent volumes (kept for up to 30 days or 1GB).

• universer: /logs/universer/
• exchange: /logs/worker-exchange/

K8s / Loki

With built-in observability, you can query logs in Grafana -> Explore using Loki:

Common Diagnostics

1. Feature outage: check SLI for RDS/MQ/Redis
2. API errors: inspect universer SLI and error codes
3. Log search: query biz_code=xxx in Loki
4. Capacity issues: monitor QPS, latency, CPU, and memory