Operations¶
Monitoring and runbooks for openframe-core in production. This section is for engineers responding to incidents or verifying system health.
Runbooks¶
| Scenario | Runbook |
|---|---|
| Adapter cannot connect to backend | Adapter Connection Failure |
| Queue consumer stops processing messages | Queue Consumer Stuck |
| OTel spans and metrics not appearing in backend | Telemetry Export Failing |
| Postgres connection pool exhausted | Postgres Pool Exhausted |
| Redis connection lost mid-operation | Redis Connection Lost |
| Modal cold starts spiking | Modal Cold Start Spike |
Recovery Tools Available¶
| Tool | Effect |
|---|---|
await port.health() |
Returns PluginHealth with status, message, and details |
registry.get(Capability.X) |
Retrieve the registered port for a capability |
modal app logs <app-name> |
Stream live logs |
modal app stop <app-name> |
Stop environment |
record_lifecycle_event("recovery", {...}) |
Record recovery event in metrics |