fix: label_replace adds its label to instant answers and to post-aggregated series, like Loki

CHANGELOG.md

@@ -7,6 +7,45 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+### Fixed
+
+- **`label_replace` adds its destination label to instant answers, reads an
+  escaped regexp, and runs under `topk`, `sort` and a binary expression (#744).**
+  `label_replace(rate({app="api-gateway"}[5m]), "app_short", "$1", "app",
+  "(.*)-.*")` came back from `/loki/api/v1/query` without `app_short` on every
+  series (also through nested `label_replace`), where Loki adds it
+  (`pkg/logql/evaluator.go` `LabelReplaceEvaluator.Next`): the proxy applied the
+  transform to range matrices only. It now rewrites the labels of a vector and
+  of a matrix alike, keeps the rest of the answer, and expands the replacement
+  the way Loki does (`$1`, `${1}`, `$name`, `$$`; an empty expansion removes the
+  destination), with the anchored regexp. Two series that end with one label
+  set stay two samples on an instant query and become one series on a range
+  query, as `JoinSampleVector` does, and no error is raised, as in Loki. Also
+  fixed: the string arguments are unquoted like the LogQL lexer does, so a
+  regexp written `"(\\w+)\\.(\\w+)"` matches (the backslashes used to be kept,
+  and the regexp never matched) and an escaped quote no longer ends the
+  argument; `topk`, `bottomk`, `sort` and `stddev` over a `label_replace` sent
+  the translator's marker to VictoriaLogs and now rank what the transform
+  produced; and a `label_replace` around a binary expression on `query_range`
+  sent the binary marker to VictoriaLogs, as the instant path used to. An
+  invalid regexp under those wrappers is Loki's `parse error : invalid regex in
+  label_replace` 400. `label_replace(sum without (level) (...), ...)` kept the
+  excluded label and lost the transform, because the `without` marker swallowed
+  the `label_replace` marker after it; and a call without the four string
+  arguments of Loki's grammar, or with an invalid escape in one, was taken for a
+  log query (an empty `streams` answer on `query_range`) and is now a 400 parse
+  error on both endpoints.
+  An instant `label_replace(count_over_time({env="production"}[5m]), ...)` (or
+  `bytes_over_time`, `rate`, ...) no longer comes back as one unlabeled total: the
+  grouping by stream was added only for a query that itself starts with the range
+  function, so the wrapper turned sixteen series into `{}`. A `label_replace`
+  around a metric over `| json` is answered by the evaluator that owns JSON parser
+  errors, as the metric is without the wrapper: a line that is not JSON fails the
+  query with Loki's `pipeline error: 'JSONParserErr'` 400 unless the pipeline
+  filters `__error__`, and `| json | __error__=""` keeps only the streams that
+  parse (the wrapper used to skip both and answer 200 with every stream). The
+  transform is applied to that answer.
+
 ## [2.13.0] - 2026-10-12
 
 ### Fixed

bench/ab/results/2026-10-12-labelreplace-1h.json

@@ -0,0 +1,232 @@
+{
+ "label": "labelreplace-1h",
+ "date": "2026-10-12",
+ "set": "label-replace",
+ "description": "label_replace adds its label to instant answers, to post-aggregated series and around a bare range function or a `| json` metric (PR for semantics/label-replace-new-label). L1 instant and L2 range are the registry row's form over a sum by; L3 is the transform under topk (ranked after the transform, as Loki evaluates it); L4 is an instant label_replace over a bare count_over_time (one series per stream, not one total); L5 and L6 are the transform over a `| json` metric that drops __error__ (answered by the evaluator that owns JSON parser errors). Run with the control set.",
+ "baseline": "base",
+ "candidate": "pr",
+ "reference": "loki",
+ "runs": 5,
+ "valid": true,
+ "restart_before": "0",
+ "restart_after": "0",
+ "noise": 0.3,
+ "timing": "warm",
+ "verdicts": {
+  "same": 5,
+  "slower": 1
+ },
+ "result_differs_from_reference": 1,
+ "result_differs_preexisting": 1,
+ "stale_answers": {
+  "base": 0,
+  "pr": 0,
+  "loki": 0
+ },
+ "rows": [
+  {
+   "shape": "L1 label_replace over sum by rate, instant",
+   "range": "instant",
+   "p50": {
+    "base": 0.005,
+    "pr": 0.005,
+    "loki": 0.109
+   },
+   "p95": {
+    "base": 0.008,
+    "pr": 0.006,
+    "loki": 0.119
+   },
+   "cold": {
+    "base": 0.014,
+    "pr": 0.005,
+    "loki": 0.099
+   },
+   "status": {
+    "base": "200",
+    "pr": "200",
+    "loki": "200"
+   },
+   "vl_cpu_max": {
+    "base": 3.49,
+    "pr": 3.49,
+    "loki": 3.49
+   },
+   "parity_vs_reference": "same",
+   "baseline_parity_vs_reference": "same",
+   "parity_vs_baseline": "same",
+   "verdict": "same",
+   "speedup": 1.0
+  },
+  {
+   "shape": "L2 label_replace over sum by rate",
+   "range": "1h",
+   "p50": {
+    "base": 0.029,
+    "pr": 0.035,
+    "loki": 0.166
+   },
+   "p95": {
+    "base": 0.032,
+    "pr": 0.037,
+    "loki": 0.187
+   },
+   "cold": {
+    "base": 0.041,
+    "pr": 0.035,
+    "loki": 0.82
+   },
+   "status": {
+    "base": "200",
+    "pr": "200",
+    "loki": "200"
+   },
+   "vl_cpu_max": {
+    "base": 29.01,
+    "pr": 29.01,
+    "loki": 29.01
+   },
+   "parity_vs_reference": "same",
+   "baseline_parity_vs_reference": "same",
+   "parity_vs_baseline": "same",
+   "verdict": "same",
+   "speedup": 0.83
+  },
+  {
+   "shape": "L3 topk over label_replace over sum by count",
+   "range": "1h",
+   "p50": {
+    "base": 0.022,
+    "pr": 0.048,
+    "loki": 0.07
+   },
+   "p95": {
+    "base": 0.023,
+    "pr": 0.053,
+    "loki": 0.141
+   },
+   "cold": {
+    "base": 0.031,
+    "pr": 0.049,
+    "loki": 0.826
+   },
+   "status": {
+    "base": "200",
+    "pr": "200",
+    "loki": "200"
+   },
+   "vl_cpu_max": {
+    "base": 39.57,
+    "pr": 40.35,
+    "loki": 40.35
+   },
+   "parity_vs_reference": "differs",
+   "baseline_parity_vs_reference": "differs",
+   "parity_vs_baseline": "differs",
+   "verdict": "same",
+   "speedup": 0.46
+  },
+  {
+   "shape": "L4 label_replace over bare count_over_time, instant",
+   "range": "instant",
+   "p50": {
+    "base": 0.003,
+    "pr": 0.007,
+    "loki": 0.013
+   },
+   "p95": {
+    "base": 0.003,
+    "pr": 0.009,
+    "loki": 0.016
+   },
+   "cold": {
+    "base": 0.003,
+    "pr": 0.006,
+    "loki": 0.012
+   },
+   "status": {
+    "base": "200",
+    "pr": "200",
+    "loki": "200"
+   },
+   "vl_cpu_max": {
+    "base": 40.35,
+    "pr": 40.35,
+    "loki": 40.35
+   },
+   "parity_vs_reference": "same",
+   "baseline_parity_vs_reference": "differs",
+   "parity_vs_baseline": "differs",
+   "verdict": "same",
+   "speedup": 0.43
+  },
+  {
+   "shape": "L5 label_replace over json sum by, instant",
+   "range": "instant",
+   "p50": {
+    "base": 0.018,
+    "pr": 0.05,
+    "loki": 0.014
+   },
+   "p95": {
+    "base": 0.022,
+    "pr": 0.056,
+    "loki": 0.015
+   },
+   "cold": {
+    "base": 0.028,
+    "pr": 0.057,
+    "loki": 0.015
+   },
+   "status": {
+    "base": "200",
+    "pr": "200",
+    "loki": "200"
+   },
+   "vl_cpu_max": {
+    "base": 36.58,
+    "pr": 36.58,
+    "loki": 36.58
+   },
+   "parity_vs_reference": "same",
+   "baseline_parity_vs_reference": "same",
+   "parity_vs_baseline": "same",
+   "verdict": "same",
+   "speedup": 0.36
+  },
+  {
+   "shape": "L6 label_replace over json sum by",
+   "range": "1h",
+   "p50": {
+    "base": 0.021,
+    "pr": 0.554,
+    "loki": 0.015
+   },
+   "p95": {
+    "base": 0.022,
+    "pr": 0.557,
+    "loki": 0.022
+   },
+   "cold": {
+    "base": 0.024,
+    "pr": 0.536,
+    "loki": 0.112
+   },
+   "status": {
+    "base": "200",
+    "pr": "200",
+    "loki": "200"
+   },
+   "vl_cpu_max": {
+    "base": 35.61,
+    "pr": 35.61,
+    "loki": 35.61
+   },
+   "parity_vs_reference": "same",
+   "baseline_parity_vs_reference": "same",
+   "parity_vs_baseline": "same",
+   "verdict": "slower",
+   "speedup": 0.04
+  }
+ ]
+}