@interopio/desktop API Reference

Configuration for the resource load instrumentation.

The phase metrics record nonzero samples only: with connection reuse most requests have no DNS/TCP/TLS phase at all, and recording those zeros would drown the distributions. Each histogram therefore answers "when this phase happened, what did it cost". Cross-origin resources without a Timing-Allow-Origin header withhold the detailed timings and sizes and contribute no phase/throughput samples.

Properties (15)

Property Type Default Required Description
cacheStateMetric⚓︎ boolean | string | null false x

If true or a string, a counter of observed resource loads by cache state is published ("io.insights.resource.cache.state" unless a string names it), tagged with cacheState ("hit" / "revalidated" / "network" / "unknown") and server.address/server.port - the cache hit ratio falls out at query time.
connectDurationMetric⚓︎ boolean | string | null false x

If true or a string, connection establishment durations (connectStart until connectEnd - includes the TLS handshake when one happened) are published as a histogram metric ("io.insights.resource.connect.duration" unless a string names it). Nonzero connects only.
connectDurationMetricBuckets⚓︎ number[] [0.001, 0.0025, 0.005, 0.01, 0.025, 0.05, 0.1, 0.25, 0.5, 1, 2.5] x

Explicit histogram bucket boundaries for connectDurationMetric, in seconds, ascending.
dnsDurationMetric⚓︎ boolean | string | null false x

If true or a string, DNS lookup durations are published as a histogram metric ("io.insights.resource.dns.duration" unless a string names it). Nonzero lookups only. Tagged with server.address/server.port and the ALPN protocol.
dnsDurationMetricBuckets⚓︎ number[] [0.001, 0.0025, 0.005, 0.01, 0.025, 0.05, 0.1, 0.25, 0.5, 1, 2.5] x

Explicit histogram bucket boundaries for dnsDurationMetric, in seconds, ascending.
downloadThroughputMetric⚓︎ boolean | string | null false x

If true or a string, effective download throughput (encodedBodySize over the responseStart-responseEnd window, in bytes per second) is published as a histogram metric ("io.insights.resource.download.throughput" unless a string names it). "network" cache state only, and only bodies of at least downloadThroughputMinBytes - smaller transfers complete within the initial congestion window and would measure round-trip time, not bandwidth. Note this is per-stream effective throughput: concurrent downloads share the link, so it reflects what a transfer experienced, not link capacity. For a byte-weighted aggregate, sum sizes and durations at query time instead of averaging rates.
downloadThroughputMetricBuckets⚓︎ number[] [65536, 262144, 1048576, 4194304, 16777216, 67108864, 268435456] x

Explicit histogram bucket boundaries for downloadThroughputMetric, in bytes per second, ascending.
downloadThroughputMinBytes⚓︎ number 32768 x

Minimum encodedBodySize, in bytes, for a resource to contribute a downloadThroughputMetric sample.
enabled⚓︎ boolean x

Whether the resource instrumentation is enabled.
firstByteDurationMetric⚓︎ boolean | string | null false x

If true or a string, resource time-to-first-byte (requestStart until responseStart) is published as a histogram metric ("io.insights.resource.ttfb.duration" unless a string names it). Measures the asset-serving infrastructure, distinct from the request instrumentation's TTFB metric which covers fetch/XHR API traffic. Samples only resources that actually hit the network ("network" and "revalidated" cache states) - cache hits would contribute disk latency, not server latency.
firstByteDurationMetricBuckets⚓︎ number[] [0.005, 0.01, 0.025, 0.05, 0.075, 0.1, 0.25, 0.5, 0.75, 1, 2.5, 5, 7.5, 10] x

Explicit histogram bucket boundaries for firstByteDurationMetric, in seconds, ascending.
skipRequests⚓︎ boolean true x

If true, resource timing entries produced by fetch/XMLHttpRequest are skipped entirely (the request instrumentation covers those requests in more detail). Set to false to also publish resource spans and metric samples for them - e.g. to include API backends in the connection-phase and throughput metrics. Beacon entries and telemetry uploads are always skipped.
tlsDurationMetric⚓︎ boolean | string | null false x

If true or a string, TLS handshake durations (secureConnectionStart until connectEnd - a subset of the connect duration) are published as a histogram metric ("io.insights.resource.tls.duration" unless a string names it). Secure connections only.
tlsDurationMetricBuckets⚓︎ number[] [0.001, 0.0025, 0.005, 0.01, 0.025, 0.05, 0.1, 0.25, 0.5, 1, 2.5] x

Explicit histogram bucket boundaries for tlsDurationMetric, in seconds, ascending.
trace⚓︎ boolean true x

If false, no resource spans are published - metrics-only mode: the phase/TTFB/throughput histograms and the cache-state counter keep reporting.