Metrics Explorer
Graph anything with PromQL or a point-and-click builder.
What it is#
A place to graph any metric Pulsimo has collected, without building a dashboard.
Where to find it#
Metrics → Explorer. Many other places open it with the query filled in: the Explore button on a source, on a chart in an endpoint's Metrics tab, or on an incident's charts.
How to use it#
- Pick a source and a metric, or type PromQL directly. Metric names autocomplete
and Tab accepts a suggestion. For counters (names ending in
_total), switch on Rate/s to see per-second rates instead of an ever-growing number. - Range: 15m, 1h, 6h, 24h, 7d, 30d, 60d, 90d (up to the retention period), or Custom for a past window, with shortcuts such as "the same day last week".
- Graph or Table (the latest value of each series). Click a legend entry to hide that series. Auto-refresh keeps a live range moving.
- Presets for quick checks: sources up, scrape duration, samples per source.
- The URL holds the query and range, so you can share a link to exactly what you see.
Useful queries#
| Question | PromQL |
|---|---|
| Requests per second per route | sum by (route) (rate(http_requests_total[5m])) |
| Share of 5xx | sum(rate(http_requests_total{status=~"5.."}[5m])) / sum(rate(http_requests_total[5m])) |
| p95 latency | histogram_quantile(0.95, sum by (le) (rate(http_request_duration_seconds_bucket[5m]))) |
| Host CPU | 1 - avg(rate(node_cpu_seconds_total{mode="idle"}[5m])) |
| Your checks, response time | pulsimo_check_duration_seconds{endpoint="Checkout API"} |
| Your checks, availability over a day | avg_over_time(pulsimo_check_up{endpoint="Checkout API"}[1d]) |
| Time spent in TLS per check | pulsimo_check_phase_seconds{phase="tls"} |
Pulsimo's own check results are metrics too (Monitoring from metrics lists them), so you can chart an endpoint's check latency next to the app's own p95.
How it works#
Every query goes through the API gateway, which adds your organization as a filter
(org_id), so you only ever see your own data. The store hides the newest ~30 s by
default, so a brand-new source appears in queries about half a minute after its first
scrape.
Limits#
- Queries older than the retention period return "outside retention".
- Heavy queries time out after 30 s; narrow the range or the label filters.