context deadline exceeded... with no indication any db connection

[zbyte64]

My sql_exporter.yml:

global:
  scrape_timeout: 1000s
  scrape_timeout_offset: 500ms
  scrape_error_drop_interval: 0s
  min_interval: 60s
  max_connections: 3
  max_idle_connections: 3
target:
  name: cloud_database
  data_source_name: '${DATABASE_URL}'
  collectors: [cloud_collector]

collectors:
  - collector_name: cloud_collector
    metrics:
      - metric_name: schedule_active
        type: gauge
        key_labels:
          - restaurant
        values: [active]
        query: |
          SELECT 
            restaurant_id AS restaurant, 
            schedule_window_active AS active
          FROM active_schedules
          WHERE enabled=true

And the log of it running using the latest published docker image:

sql_exporter-1  | time=2026-02-09T18:54:43.803Z level=WARN source=main.go:99 msg="Starting SQL exporter" versionInfo="(version=0.19.0, branch=HEAD, revision=763350af5055e002eaa02c53f2e634b94e78c9f4)" buildContext="(go=go1.25.6, platform=linux/amd64, user=root@buildkitsandbox, date=20260202-15:37:14, tags=netgo)"
sql_exporter-1  | time=2026-02-09T18:54:43.804Z level=WARN source=collector.go:67 msg="Non-zero min_interval, using cached collector." logContext="collector=cloud_collector" min_interval=1m
sql_exporter-1  | time=2026-02-09T18:54:43.805Z level=INFO source=tls_config.go:354 msg="Listening on" address=[::]:9399
sql_exporter-1  | time=2026-02-09T18:54:43.806Z level=INFO source=tls_config.go:357 msg="TLS is disabled." http2=false address=[::]:9399
sql_exporter-1  | time=2026-02-09T18:55:22.536Z level=ERROR source=promhttp.go:55 msg="Error gathering metrics" error="[from Gatherer #1] context deadline exceeded"
sql_exporter-1  | time=2026-02-09T18:55:22.536Z level=ERROR source=promhttp.go:58 msg="No metrics gathered" error="[from Gatherer #1] context deadline exceeded"
sql_exporter-1  | time=2026-02-09T18:56:22.535Z level=ERROR source=promhttp.go:55 msg="Error gathering metrics" error="[from Gatherer #1] context deadline exceeded"
sql_exporter-1  | time=2026-02-09T18:56:22.535Z level=ERROR source=promhttp.go:58 msg="No metrics gathered" error="[from Gatherer #1] context deadline exceeded"

The "context deadline" message seems to appear every minute when alloy goes to collect metrics.

Notable I don't see any "Database handle successfully"; It seems that if I change my credentials to something I know to be invalid I get the same results, but I would expect to see an auth error instead (like when I specify an incorrect driver).

Is there a verbosity setting that could tell me how many metrics it's going to collect and when it attempts to do so, or when a db connection is made?

Edit: Updated with output from 0.19.0

[burningalchemist]

Hi @zbyte64, I noticed that you're using a version v0.5 from 2019, which isn't supported.

Besides, sql_exporter doesn't perform scrapes on its own, the queries are executed when your collector hits the endpoint of sql_exporter.

Generally, 'context deadline exceeded' error means that the request couldn't succeed within a defined timeout, which might imply network issues, connection parameters misconfiguration, etc.

I'd suggest updating to the latest version (e.g. v0.19.0) first. There's been a lot of changes and improvements since that time.

UPD: Just to clarify, the original repo is unmaintained for a long time, I develop and maintain its fork and successor since 2019, so might need to change the image - please check the docs. 👍

[zbyte64]

Woops, I had tried both repositories to see if I get different results and pasted the older one. Here it is with 0.19.0:

sql_exporter-1  | time=2026-02-09T18:54:43.803Z level=WARN source=main.go:99 msg="Starting SQL exporter" versionInfo="(version=0.19.0, branch=HEAD, revision=763350af5055e002eaa02c53f2e634b94e78c9f4)" buildContext="(go=go1.25.6, platform=linux/amd64, user=root@buildkitsandbox, date=20260202-15:37:14, tags=netgo)"
sql_exporter-1  | time=2026-02-09T18:54:43.804Z level=WARN source=collector.go:67 msg="Non-zero min_interval, using cached collector." logContext="collector=cloud_collector" min_interval=1m
sql_exporter-1  | time=2026-02-09T18:54:43.805Z level=INFO source=tls_config.go:354 msg="Listening on" address=[::]:9399
sql_exporter-1  | time=2026-02-09T18:54:43.806Z level=INFO source=tls_config.go:357 msg="TLS is disabled." http2=false address=[::]:9399
sql_exporter-1  | time=2026-02-09T18:55:22.536Z level=ERROR source=promhttp.go:55 msg="Error gathering metrics" error="[from Gatherer #1] context deadline exceeded"
sql_exporter-1  | time=2026-02-09T18:55:22.536Z level=ERROR source=promhttp.go:58 msg="No metrics gathered" error="[from Gatherer #1] context deadline exceeded"

[burningalchemist]

Sounds good, thanks!

Could you please share your configuration, specifically the connection string? Please make sure to redact the sensitive data, replace it with some random chars.

For verbosity, please try --log.level=debug.

[zbyte64]

The DATABSE_URL looks like:

postgresql://user:pass@site-database-1.cluster-abcdefg512345.us-region-1.rds.amazonaws.com:5432/db_name

[burningalchemist]

I'd recommend checking the following:

• Since you're on AWS, check the security groups for the open port 5432. I can't exclude this, so if you've already checked and you're sure you can connect to the instance with any other app, then feel free to ignore;
• if the password contains any special characters, we need to encode them (please check https://github.com/burningalchemist/sql_exporter?tab=readme-ov-file#database-urls-url-format-dsns for more details);
• check the scrape_interval on the sql_exporter side, but also on the alloy (or any other collector) side. min_interval is purposed for caching (to avoid hitting DBs too much);
• The connection string expansion from the config might not work correctly, so instead try using SQLEXPORTER_TARGET_DSN environment variable. It will override the parameter from the config.

(Although, if your custom environment variable isn't expanded correctly, it should appear as an error in the logs. It's possible you may need to try setting $DATABASE_URL in the config without curly braces)

Please try the above and let me know. 😃👍