UseSlidingInvisibilityTimeout not working sometimes

[strigefleur]

Having configuration

InvisibilityTimeout = TimeSpan.FromMinutes(5),
UseSlidingInvisibilityTimeout = true,

I'd expect it to let long running jobs to execute without setting huge InvisibilityTimeout.

However under some circumstances (alas, can't reproduce intentionally) - like for a week it is fine and than just no - the task is interrupted after 5 minutes (InvisibilityTimeout I guess) with OperationCanceledException and then is tried to requeue (again sometimes there's also a warning about another worker that fetches the task).

Any ideas?

[strigefleur]

Looking at the source It has to be something with worker change

        string updateFetchAtSql = $@"
          UPDATE ""{_storage.Options.SchemaName}"".""jobqueue"" 
          SET ""fetchedat"" = NOW()
          WHERE ""id"" = @id AND ""fetchedat"" = @fetchedAt
          RETURNING ""fetchedat"" AS ""FetchedAt"";
        ";

        try
        {
          _storage.UseConnection(null, connection =>
          {
            FetchedAt = connection.ExecuteScalar<DateTime?>(updateFetchAtSql,
              new { queue = Queue, id = Id, fetchedAt = FetchedAt });
          });

          if (!FetchedAt.HasValue)
          {
            _logger.Warn(
              $"Background job identifier '{JobId}' was fetched by another worker, will not execute keep alive.");
          }

so that when fetchedat has changed externally there's.. what? something's wrong?
Why would the worker switch anyway?

[latop2604]

Hello, I also have some issue with InvisibilityTimeout.

I set it to 6 hours. But, after 30 minutes, the job is flagged as cancelled and is taken by an other worker, while the initial one is still processing.

I don't use UseSlidingInvisibilityTimeout and have WorkerCount set to 1.

Version used

• Hangfire.AspNetCore 1.8.18
• Hangfire.PostgreSql 1.20.12

We used to have very long running job. And by setting InvisibilityTimeout it was working fine previously