Monitoring Tool
Know which endpoint is failing, and how badly, before your users tell you.
Your API sends each request to the ingest endpoint, one at a time or in batches of up to 100. Monitoring Tool groups them by route and minute and shows requests, errors and latency for every endpoint.
The demo is shared and read-only. Its data is synthetic and refreshed every hour.
Requests per minute; the darker part is 5xx errors. Point at a bar to see that minute.
| Endpoint | Req / min | 5xx | p95 |
|---|---|---|---|
| GET /api/designs | 535 | 0.00% | 42 ms |
| GET /s/{id} | 401 | 0.00% | 21 ms |
| GET /api/surfaces | 161 | 0.00% | 32 ms |
| POST /api/share | 134 | 0.00% | 59 ms |
| POST /api/uploads | 107 | 0.00% | 897 ms |
Features
What it does
Every endpoint at a glance
Requests are grouped by route, minute, status class and latency bucket. IDs in paths are grouped for you, so /users/42 and /users/43 count as one route.
SLO alerts
Set availability and latency objectives. Multi-window burn-rate alerts, threshold rules and no-traffic rules send an email or a signed webhook.
Private by default
Request bodies are not stored unless you turn it on. Sensitive headers are masked and IP addresses shortened before storage.
How it works
From one request to one alert
Alerts use multi-window burn rates, the method in Google's SRE Workbook, so a fast burn and a slow leak both reach you and a single spike does not.