Usage
Usage tells you how hard you’re working Saiku Cloud. Four headline numbers, three charts, and a date range. It defaults to the last 30 days.
Come here when you want to know whether you’re heading for a rate-limit ceiling, whether last night’s error spike was tenant-wide or one bad integration, or whether your file storage is creeping toward your plan’s cap.
The headline numbers
Total events. Every request in the range.
Error rate. The share of requests that returned HTTP 400 or above, with the raw counts underneath so you can tell 0.4% of thirteen thousand from 0.4% of thirty. Healthy steady state is well under 1%; a sustained climb is worth chasing in the audit log, which can tell you which paths and which actors.
P95 latency, with P50 and P99 alongside it. P50 is what a typical user feels; P95 is the number to quote in an SLA conversation; P99 is the pathological-query story.
File storage. Current usage against your plan’s cap, as a percentage and a figure. Same number the Files page shows, repeated here so it’s visible next to everything else.
The charts
Three of them, all over the same range: events per bucket, latency percentiles (P50 / P95 / P99 over time), and error rate as a percentage.
Choosing a range
Set From and To, pick a bucket, and click Apply. Nothing re-queries as you type — on a busy tenant that would be unusable — and the range lands in the URL, so you can bookmark a view or paste it to a colleague and know they’ll see what you saw.
Bucket is Auto by default: hourly for ranges of a week or less, daily beyond that. You can force one, with one restriction — hourly buckets are refused on ranges longer than seven days, because 30 days of hourly points is 720 bars nobody can read. The maximum range is 90 days.
The numbers are read live from the same audit trail the Audit log shows. There’s no pipeline in between and no lag to wait out — reload and you have the current picture.
What you won’t see here
- Per-cube breakdown. This is a tenant-wide rollup. The audit log has per-event detail if you need to slice by cube.
- Per-user breakdown. Same answer — the audit log is the route to per-actor visibility.