PromQL: parser has wrong precedence with unary + and -, affecting offset arithmetic #16711
Description
What did you do?
Started prometheus compiled from 7370d62acfcbf08e0528c81857d165ee6ca1940b with or without the --enable-feature=promql-duration-expr feature flag and queried variants of the following:
curl -XPOST localhost:9090/api/v1/query --data-urlencode 'query=foo offset -1m * bar'
What did you expect to see?
I did expect to see no parse errors.
What did you see instead? Under which circumstances?
I did see a parse error:
{"status":"error","errorType":"bad_data","error":"invalid parameter \"query\": 1:18: parse error: unexpected identifier \"bar\" in offset, expected number or duration"}
Its even worse, if i toggle the featureflag and query something that the parser does parse but parses wrong:
curl -XPOST localhost:9090/api/v1/query --data-urlencode 'query=up offset -1m * 100'
{"status":"success","data":{"resultType":"vector","result":[]}}
vs
curl -XPOST localhost:9090/api/v1/query --data-urlencode 'query=up offset (-1m) * 100'
{"status":"success","data":{"resultType":"vector","result":[{"metric":{"app":"prometheus","instance":"localhost:9090","job":"prometheus"},"value":[1749559998.482,"100"]}]}}
"up offset -1m * 100" seems to be parsed as "up offset (-1m * 100)" instead of "(up offset -1m) * 100"
System information
No response
Prometheus version
7370d62acfcbf08e0528c81857d165ee6ca1940b
Prometheus configuration file
Just used the example configuration from the repository:
# my global config
global:
scrape_interval: 15s # Set the scrape interval to every 15 seconds. Default is every 1 minute.
evaluation_interval: 15s # Evaluate rules every 15 seconds. The default is every 1 minute.
# scrape_timeout is set to the global default (10s).
# Alertmanager configuration
alerting:
alertmanagers:
- static_configs:
- targets:
# - alertmanager:9093
# Load rules once and periodically evaluate them according to the global 'evaluation_interval'.
rule_files:
# - "first_rules.yml"
# - "second_rules.yml"
# A scrape configuration containing exactly one endpoint to scrape:
# Here it's Prometheus itself.
scrape_configs:
# The job name is added as a label `job=<job_name>` to any timeseries scraped from this config.
- job_name: "prometheus"
# metrics_path defaults to '/metrics'
# scheme defaults to 'http'.
static_configs:
- targets: ["localhost:9090"]
# The label name is added as a label `label_name=<label_value>` to any timeseries scraped from this config.
labels:
app: "prometheus"