Error monitoring

[jtourkos]

When an event fails after all retries, we just log it — no alerts, no visibility. That makes it hard to:

• Spot recurring failure patterns
• Catch critical issues quickly
• Tell recoverable errors from unrecoverable ones

Current behavior:

• 20 retries, 30-second backoff (saveEventProcessingJob.ts:35-36)
• Logs look like: ❌ [${job}] failed: ${err.message} (queue.ts:35)
• No error categorization
• No alerting for critical failures

We need to find a way to be notified about (critical) errors.

[efstajas]

• add opentelemetry sdk, connect to grafana collector
• create traces for every job processing
• ensure traces have error state if failed
• add postgres auto-instrumentation
• set up alert on grafana

[mhgbrown]

@jtourkos is this covered now in Dripfeed?