Hello,
We have collect a Prometheus counter called http_request_total. The counter has three labels, http_status (e.g. 200, 300, 400, 500), endpoint (e.g. /people/ryan), and method (e.g. GET, POST). I’m trying to create a dashboard that can easily show me the number of 200s, 300s, 400s, and 500s per method and endpoint. I thought that the Table panel would be the best to display this information, but I’m struggling to get everything to work. I’d like the table to look like the following:
| endpoint | method | 200s | 300s | 400s | 500s |
|---|---|---|---|---|---|
| /users | POST | 1000 | 5 | 7 | 10 |
| /users | GET | 5000 | 0 | 0 | 0 |
The time range that was being queried would be controlled by the Grafana UI (e.g. a “Quick Range” like “Last 1 hour”). Is this possible?
So far, the best I’ve been able to do is to get a single HTTP status to aggregate per table using:
- Metrics
- Query: sum(web_api_requests_total{http_status=~“2…”}) by (endpoint, method)
- Legend format: {{endpoint}} {{method}}
- Format as: Time Series
- Options:
- Table Transform: Time series aggregation
- Columns: Total
That gives me a table that looks like:
| Metric | Total |
|---|---|
| /users POST | 1000 |
| /users GET | 5000 |
Does anyone have any suggestions on how best to display a sum of HTTP status codes over the given time period by endpoint and method?
Thanks,
Ryan