Context deadline exceeded error handling

[Maros112358]

Describe the bug
When database timeouts, error log on line 40 does not show in the logs, even though error log on line 47 shows a list of errors, where one of them is context deadline exceeded.

To Reproduce
Steps to reproduce the behavior:

1. Use the config below
2. Go to /metrics
3. Check logs
4. See error

Expected behavior
I expected to see Timeout collecting metrics from target: ...

Configuration

jobs:
  - job_name: my_job
    collectors: [mssql_timeout]
    static_configs:
      - targets:
          'my_target': 'my_connection_string'
collectors:
  - collector_name: mssql_timeout
    metrics:
      - metric_name: timeout_metrics
        type: gauge
        help: 'wait for 3 minutes in MSSQL'
        values: [xxx]
        query: |
          WAITFOR DELAY '00:03:00';

Additional context
Logs I got when reproducing the error:

ts=2023-10-02T12:04:36.730Z caller=klog.go:108 level=warn func=Warningf msg="Starting SQL exporter (version=0.12.3, branch=HEAD, revision=d25cd797ff6b4daa53a5b9cc29c4aee3c998cb9f) (go=go1.20.8, platform=darwin/arm64, user=root@3c5818fec6b5, date=20230920-16:18:01, tags=netgo)"
ts=2023-10-02T12:04:36.737Z caller=tls_config.go:274 level=info msg="Listening on" address=[::]:9399
ts=2023-10-02T12:04:36.737Z caller=tls_config.go:277 level=info msg="TLS is disabled." http2=false address=[::]:9399
ts=2023-10-02T12:04:49.060Z caller=klog.go:124 level=error func=Errorf msg="Error gathering metrics: [from Gatherer #1] [job=\"my_job\", target=\"my_target\", collector=\"mssql_timeout\", query=\"timeout_metrics\"] context deadline exceeded"

[burningalchemist]

Hey @Maros112358 thanks for reporting, I'll have a look. 👍

[Maros112358]

I removed the if statement and experimented with the code, but the following code does not give any more information than Error gathering metrics:

				klog.Errorf("Timeout collecting metrics from target: %s", err.Error())

It may be the problem of the underlying driver.

[Maros112358]

I just found out, if you try to reproduce the bug, the sql_exporter also caches result of a "timeout" query.

How to reproduce:
0, use the aforementioned config
1, set min_interval to 1m
2, go to /metrics
3, wait for timeout → scrape_duration 10 seconds + context deadline exceeded
4, refresh the page → scrape duration under 1 second + context deadline exceeded

Is this desired behaviour?

[burningalchemist]

Hey @Maros112358, thanks a lot for the test case, I'll reproduce and fix it on my end. 👍