Timeout while collecting metrics

[shukui-ipo]

Hello, I have just started using sql_exporter. I added a custom collector to query index fragmentation. Because this collector takes a long time to run, I configured “min_interval: 12h”. My Prometheus scrapes metrics every 15 seconds. However, I observed that each time Prometheus scrapes, sql_exporter prints the following error:

time=2026-01-30T10:17:11.593+08:00 level=ERROR source=promhttp.go:49 msg="Timeout while collecting metrics" error="[from Gatherer #1] [job=sqlserver_targets_indexes,target=db3,collector=mssql_indexes] context deadline exceeded"

In addition, the database also executes the query once for each scrape.

global:
  scrape_timeout: 1000s
  scrape_timeout_offset: 500ms
  max_connections: 10
  max_idle_connections: 10
  min_interval: 10s

jobs:
  - job_name: sqlserver_targets_indexes
    collectors: [mssql_indexes]
    enable_ping: true
    static_configs:
      - targets:
          db3: "sqlserver://"
        labels:
          cluster: DB
collectors:
 - collector_name: mssql_indexes
    min_interval: 12h
    metrics:
          ...........

[burningalchemist]

Hi @shukui-ipo, sql_exporter doesn't scrape targets on its own, the scrapes are controlled and fully performed by the Prometheus instance. In other words, if Prometheus doesn't hit the /metrics endpoint of sql_exporter, nothing is scraped.

If you scrape metrics from Prometheus every 15s, this literally means you're executing queries on the SQL Server every 15s. The timeouts messages you're hitting are related to that and possibly to scrape_timeout which you've set to 1000s which is 16m.

Please check the documentation section on the scrape_timeout:

# Scrape timeouts ensure that:
  #   (i)  scraping completes in reasonable time and
  #   (ii) slow queries are canceled early when the database is already under heavy load
  # Prometheus informs targets of its own scrape timeout (via the "X-Prometheus-Scrape-Timeout-Seconds" request header)
  # so the actual timeout is computed as:
  #   min(scrape_timeout, X-Prometheus-Scrape-Timeout-Seconds - scrape_timeout_offset)
  #
  # If scrape_timeout <= 0, no timeout is set unless Prometheus provides one. The default is 10s.

min_interval (either set globally or per collector configuration) makes sql_exporter cache the last successful scraped result for a given interval, which in turn protects queried endpoints from excessive usage.

---

As for your problem, it depends on how long it takes for your queries to run. The options I see:

• Depending on the query duration, you might need to configure scraping on both Prometheus and sql_exporter sides to match. It makes sense if query duration is reasonable (i.e. not taking hours);
• You can create a stored procedure on the MSSQL side that queries and writes/caches the result to a table on a SQL Server side, so then you can collect this value instantly using sql_exporter;
• You can try alternative exporters that execute queries asynchronously (please check README for more information).