API Monitoring

Monitor API and service health with flat numeric events. Keep events specific — one event per signal, not one event per endpoint.

Recommended events

Event	Value	Why
`api.requests.count`	`1` per request	Track total request volume
`api.errors.count`	`1` per error	Catch error rate increases
`api.response.duration_ms`	Duration in ms	Monitor response time and tail latency
`rate_limit.hit.count`	`1` per hit	Detect when clients are being throttled

Logging events

JavaScript


await client.log({ agent: "api.v1", event: "api.requests.count", value: 1,
  meta: { method: "POST", path: "/v1/orders" } });
 
await client.log({ agent: "api.v1", event: "api.response.duration_ms", value: 145,
  meta: { method: "POST", path: "/v1/orders", status: 201 } });
 
await client.log({ agent: "api.v1", event: "api.errors.count", value: 1,
  meta: { method: "GET", path: "/v1/users", status: 500, error: "db_timeout" } });

Python


client.log(Log(agent="api.v1", event="api.requests.count", value=1,
    meta={"method": "POST", "path": "/v1/orders"}))
 
client.log(Log(agent="api.v1", event="api.response.duration_ms", value=145,
    meta={"method": "POST", "path": "/v1/orders", "status": 201}))
 
client.log(Log(agent="api.v1", event="api.errors.count", value=1,
    meta={"method": "GET", "path": "/v1/users", "status": 500, "error": "db_timeout"}))

Go


_ = client.Log(ctx, chirpier.Log{Agent: "api.v1", Event: "api.requests.count", Value: 1,
    Meta: map[string]any{"method": "POST", "path": "/v1/orders"}})
 
_ = client.Log(ctx, chirpier.Log{Agent: "api.v1", Event: "api.response.duration_ms", Value: 145,
    Meta: map[string]any{"method": "POST", "path": "/v1/orders", "status": 201}})
 
_ = client.Log(ctx, chirpier.Log{Agent: "api.v1", Event: "api.errors.count", Value: 1,
    Meta: map[string]any{"method": "GET", "path": "/v1/users", "status": 500, "error": "db_timeout"}})

Recommended first policies

Event	Period	Aggregate	Condition	Threshold	Why
`api.errors.count`	`hour`	`sum`	`gt`	`50`	Catch sustained error spikes
`api.response.duration_ms`	`hour`	`p95_est`	`gt`	`2000`	Tail latency affecting real users
`api.response.duration_ms`	`hour`	`average`	`gt`	`500`	General slowdown detection
`rate_limit.hit.count`	`hour`	`sum`	`gt`	`100`	Clients being throttled too often

Using comparisons

Compare this hour’s error count to the same hour yesterday to distinguish a real incident from normal patterns:


const analytics = await client.getEventAnalytics(errorsEventID, {
  view: "window", period: "1h", previous: "previous_1d",
});
// If value_pct_change is +200%, errors tripled vs the same hour yesterday — investigate
// If value_pct_change is ~0%, this is baseline noise

Compare today’s request volume to the same day last week to spot traffic drops:


const analytics = await client.getEventAnalytics(requestsEventID, {
  view: "window", period: "1d", previous: "previous_7d",
});
// A -50% drop in count_pct_change might mean a routing issue