Prometheus Exposition — Broken Histograms, Intentionally Invalid (prom)
Two histogram families that are syntactically fine and semantically wrong: one has non-monotonic buckets and no +Inf bucket, the other reports a _count larger than its +Inf bucket. Intentionally invalid, because histogram_quantile over either produces a confidently wrong number rather than an error.
# HELP broken_histogram_seconds Buckets out of order with no +Inf bucket.
# TYPE broken_histogram_seconds histogram
broken_histogram_seconds_bucket{le="0.1"} 900
broken_histogram_seconds_bucket{le="0.05"} 1200
broken_histogram_seconds_bucket{le="1"} 400
broken_histogram_seconds_sum 88.5
broken_histogram_seconds_count 2400
# HELP count_disagrees_seconds _count does not equal the +Inf bucket.
# TYPE count_disagrees_seconds histogram
count_disagrees_seconds_bucket{le="0.5"} 100
count_disagrees_seconds_bucket{le="+Inf"} 150
count_disagrees_seconds_sum 12.25
count_disagrees_seconds_count 175
Specifications
- Families
- 2
- Non Monotonic Buckets
- true
- Missing Inf Bucket
- true
- Count Disagrees With Inf
- true
- Intentionally Invalid
- true
Testing contract
Expected to fail- Scenario
- Validate histogram invariants before computing any quantile.
- Expected result
- Both families are rejected — non-monotonic bucket counts and a missing +Inf bucket in the first, _count 175 against a 150 +Inf bucket in the second — instead of yielding a plausible-looking percentile.
What is a .prom file?
A .prom file holds metrics in the Prometheus text exposition format, the same body a `/metrics` endpoint returns. Each metric family is introduced by `# HELP` and `# TYPE` comments and followed by one sample per line: a metric name, an optional brace-delimited label set, a value, and an optional millisecond timestamp. Counters, gauges, histograms (with `_bucket`, `_sum`, `_count` series and a `+Inf` bucket), and summaries with quantile labels all use this one grammar.
How to use this file
Use an example .prom file to test exposition-format parsers, scrapers, and the node_exporter textfile collector — checking label escaping, histogram bucket ordering and cumulative counts, and the handling of `NaN` and `+Inf` values.
How to use this file for testing
“Prometheus Exposition — Broken Histograms, Intentionally Invalid (prom)” is a deterministic Novus Examples fixture for Observability, Schema validation, Error handling. Structured and plain-text telemetry with known timestamps, levels, request identifiers, and error states for testing log ingestion, correlation, dashboards, and alert pipelines.
Documented properties for this file: PROM · 594 bytes. Compare results against paired or grouped companions on this page when present (clean↔damaged, searchable↔scanned, or format twins) so scores stay reproducible across runs.
Download the file once, keep the path stable in CI or local scripts, and treat the spec table as the contract: dimensions, seeds, field lists, and roles are intentional. Corrupt or invalid samples are labelled as such — expect parsers to fail loudly rather than silently accept them.
Telemetry fixtures use fixed trace IDs, span IDs, and timestamps so ingestion is reproducible run to run. Point your collector, parser, or query layer at the file and assert the documented span tree, metric families, or severity mix; service and host names are invented.
Related files
- promOpenMetrics — Missing # EOF Marker, Intentionally Invalid (prom)A document that parses perfectly as Prometheus text and is invalid as OpenMetrics, because the mandatory # EOF terminator is missing. That marker is the format's only defence against a truncated response being accepted as complete — which is the exact failure this file reproduces.
- jsonOTLP Trace — All-Zero Trace and Span IDs, Intentionally Invalid (json)A trace whose trace ID and root span ID are all zeroes — the value an uninitialised context produces. Intentionally invalid: the specification defines all-zero IDs as invalid, so a conformant receiver must reject the spans instead of creating a permanent bucket that every broken service writes into.
- jsonOTLP Trace — Duplicate Span IDs, Intentionally Invalid (json)Two different spans share one span ID inside a single trace, which the specification forbids. Intentionally invalid: it is the fixture for the de-duplication path, where a naive store silently overwrites one span with the other and loses a subtree.
- jsonOTLP Trace — Negative and Zero Durations, Intentionally Invalid (json)One span ends 8.4 ms before it started and another has identical start and end timestamps. Intentionally invalid: negative durations poison latency aggregates — a single one can drag a p99 below zero — so this is the fixture for the validation that must sit in front of any duration metric.
- txtW3C traceparent — Valid and Invalid Header Cases (txt)A labelled set of traceparent header values: four that must parse — including a future version with extra fields, which parsers are required to accept — and eight that must be rejected, covering all-zero IDs, version ff, short and uppercase trace IDs, and missing or trailing fields.
- promOpenMetrics — Full Document with UNIT, info and stateset (prom)A complete OpenMetrics document — the _total and _created series a counter really has, UNIT metadata, an info metric carrying build metadata and a stateset with exactly one active state — terminated by the mandatory # EOF. This is what the exposition format became once it was standardised.
Generated by generation/observability.py. Free for any use, no attribution required — license.