query returned wrong number of labels: 1, 0

[wenerme]

• exporter
  • https://github.com/wrouesnel/postgres_exporter
  • node_exporter
• one timescale-prometheus - docker installed - latest
• tow prometheus
  • prom-a one for collect
  • prom-b one for read by grafana

but query process_start_time_seconds will failed on prom-b, prom-a can execute correctly

remote_read: remote server http://10.1.1.3:9202/read returned HTTP status 500 Internal Server Error: query returned wrong number of labels: 1, 0

https://github.com/timescale/timescale-prometheus/blob/6693256251629559a609f9e7f56131d6107eba2f/pkg/pgmodel/sql_reader.go#L252

both node_exporter and postgres_exporter has process_start_time_seconds metric

error log

{
  "caller": "read.go:45",
  "err": "query returned wrong number of labels: 1, 0",
  "level": "warn",
  "msg": "Error executing query",
  "query": {
    "queries": [
      {
        "start_timestamp_ms": 1596877133800,
        "end_timestamp_ms": 1596881033800,
        "matchers": [
          { "name": "__name__", "value": "process_start_time_seconds" }
        ],
        "hints": {
          "step_ms": 14000,
          "start_ms": 1596877133800,
          "end_ms": 1596881033800
        }
      }
    ]
  },
  "storage": "PostgreSQL",
  "ts": "2020-08-08T10:03:53.854Z"
}

[JLockerman]

This should be fixed by PR #184

[cevian]

@wenerme This should be fixed in release 0.1.0-beta.2. Can you please confirm?

[wenerme]

Works now.