Just switched from sysadmin to DevOps and started setting up our monitoring. We use Grafana dashboards for everything now – servers, apps, even some business metrics.
But our finance team heard about this and wanted a "single pane of glass" for expenses too. They forced our simple expense approval tool to pipe data into our observability stack. Now we have Prometheus metrics for expense report counts and a Grafana dashboard next to server CPU. It's a mess.
```yaml
# Example of what they asked for in a recording rule
- record: job:expense_reports_submitted:rate5m
expr: rate(expense_submissions_total[5m])
```
The dashboard has alerts when submission rates drop. But it doesn't help with actual approvals or receipts. The old tool was simple. Now accountants log into Grafana and are confused by terms like "rate" and "vector." Sync to the accounting system fails silently sometimes, and the graphs don't help fix it.
Anyone else see this? Choosing one platform (like our observability suite) for everything can add more moving parts, not simplify. Maybe some tools should just do their own job well.