Data sources

Every query reads one source. A source has measures (what you compute) and fields (what you filter, group and build variables on).

The editor and the API read the same list. GET /dashboards/catalog returns it, so the tables below always match what the query builder offers.

How to read the tables

• Aggregations are the ways a measure can be combined across a bucket and across a window. The first one listed is the default. A measure with no list has one fixed meaning.
• Group by means you can split a query by the field, up to two fields per query.
• Variable means a variable can be bound to the field.
• Kind is id (values have labels, such as monitor names), enum, text or number.
• A field is filter only when it supports neither.
• Aggregations: avg, sum, min and max; the percentiles p50, p75, p90, p95 and p99; and rate for the metrics source.
• Every query is scoped to your account.

Monitors

monitors reads HTTP and browser checks. Uptime and response time are measured per check, so they match the monitor page.

Over two days or more, widgets read hourly summaries of the checks, so long ranges stay fast. Counts, uptime, averages, min and max are exact; percentiles are within 1%.

Measure	Unit	Aggregations	Meaning
uptime	percent	fixed	Share of checks that passed.
response_time	ms	avg, p50, p75, p90, p95, p99, min, max	Response time of a check.
checks	count	fixed	Checks run.
failed_checks	count	fixed	Checks that failed.
dns_time, connect_time, tls_time, ttfb, download_time	ms	avg, p50, p95, p99	The phases of an HTTP check. HTTP monitors only.
downtime	seconds	sum	Time spent failing, derived from state transitions. Includes cron monitors.

Field	Kind	Group by	Variable
monitor	id (shown by name)	Yes	Yes
monitor_type	enum: http, browser, cron	Yes	Yes
tag	enum	Yes	Yes
region	enum	Yes	Yes
status_page	id (shown by name)	No	Yes
service	id (shown by name)	Yes	Yes

Buckets of 60 seconds and up.

Incidents

incidents reads incidents, including those opened automatically by monitors.

Measure	Unit	Aggregations	Meaning
incidents	count	fixed	Incidents created in the bucket.
open_incidents	count	fixed	Incidents open now. A single value only, never a series.
mttr	seconds	avg, p50, p90, max	Time to resolve, for incidents resolved in the window.
mtta	seconds	avg, p50, p90, max	Time to acknowledge, for incidents acknowledged in the window.

Field	Kind	Group by	Variable
severity	enum	Yes	Yes
status	enum	Yes	Yes
service	id	Yes	Yes
status_page	id	Yes	Yes
auto_created	enum	Yes	No

Buckets of 300 seconds and up.

RUM

rum reads real user monitoring data. It only sees the RUM applications in your account.

Measure	Unit	Aggregations	Meaning
sessions	count	fixed	Sessions.
views	count	fixed	Page views.
errors	count	fixed	Errors.
error_session_rate	percent	fixed	Share of sessions with an error.
lcp, inp, fcp, ttfb, load_time	ms	p75, p50, p90, p95, p99, avg	Web Vitals and load time.
cls	score	p75, p50, p90, p95, p99, avg	Cumulative layout shift.
rage_clicks	count	fixed	Rage clicks.

Field	Kind	Group by	Variable
app	id (shown by name)	Yes	Yes
route	text	Yes	Yes
url	text	No	No
country, browser, os, device_type, env, release, service	enum	Yes	Yes

Buckets of 300, 3600, 21600 and 86400 seconds.

Logs

logs counts log events.

Measure	Unit	Aggregations	Meaning
logs	count	fixed	Log events.

Field	Kind	Group by	Variable
level, source, env, host, version	enum	Yes	Yes
message	text	No	No
tag:<key>	tag value	Yes	Yes

message supports contains only. A tag field is written tag: plus the tag key, and the key can use A-Z, a-z, 0-9, ., _, - and /, up to 128 characters.

Buckets of 60 seconds and up.

Metrics

metrics reads any metric you send. Every query sets metric, the metric name. The name can also be a variable, such as $metric.

Measure	Unit	Aggregations	Meaning
value	the metric’s own unit	avg, sum, min, max, p50, p75, p90, p95, p99, rate	The metric’s value per bucket.

Field	Group by	Variable
host, env, source	Yes	Yes
tag:<key>	Yes	Yes

For what each aggregation means on gauges and counters, see the metrics overview.

Buckets of 60, 300, 3600 and 21600 seconds. One-day buckets are built from the 21600-second ones.

Traces

traces reads spans. Requests are counted as entry spans.

Measure	Unit	Aggregations	Meaning
requests	count	fixed	Entry spans.
errors	count	fixed	Errors.
error_rate	percent	fixed	Error percentage.
latency	ms	avg, p50, p75, p90, p95, p99, max	Latency.

Field	Group by	Variable
service, operation	Yes	Yes
status, span_kind	Yes	No

status and span_kind are enums.

Buckets of 60 seconds and up.

Choosing values

Dropdowns for filters and variables load their options from GET /dashboards/values. It returns at most 100 values, matched by prefix or substring as you type. Options for RUM, logs, metrics and traces come from the last 7 days. Fields with id kind show names, such as monitor and RUM application names.

Buckets

A time chart groups points into buckets. Site Qwality picks the bucket size for you.

• Ladder. Bucket sizes are 60 seconds, 5 minutes, 15 minutes, 1 hour, 6 hours and 1 day.
• Choice. The widget asks for about its pixel width divided by 4 points, and the API clamps that to between 30 and 600 (240 by default). The bucket is the smallest ladder step that gives no more points than that and that the source supports.
• Alignment. One-day buckets start at midnight in the timezone the dashboard is showing. Smaller buckets align in UTC.
• Partial bucket. A bucket that ends after now is still filling. Charts shade it.
• Scalars. Query value, Top list, Table and Pie reduce over the whole window with the measure’s aggregation. They never take the last bucket.
• Longest window. 400 days.

See also

• Queries, formulas and variables
• Widgets
• Metrics, Logs, Traces and RUM