I have an App Service which hosts 6 web-jobs. Here, one web-job has CosmosDBTrigger and my expectation is that it gets triggered every time there is new change feed in the container.

I am noticing, 8/10 times, that this feed trigger is taking almost an hour. Meaning, when the record is added to the container, I see it takes upto an hour to send this feed to this webjob.

How can I improve this? What could be the root cause?

Web-Job code:

Listener.cs:

public async Task ProcessMessageAsync(
[CosmosDBTrigger(
    "%DatabaseName%", "%ContainerName%", 
    Connection = Configurations.CosmosDbConnectionString, 
    CreateLeaseContainerIfNotExists = true, 
    LeaseContainerName = "leases", 
    LeaseContainerPrefix = nameof(EventsListener))]
    string documents, CancellationToken cancellationToken)
{
}

Program.cs

var builder = new HostBuilder()
    .UseEnvironment(environment)
    .ConfigureWebJobs(b =>
    {
        b.AddCosmosDB(configDbOptions =>
        {
            configDbOptions.ConnectionMode = ConnectionMode.Gateway;
        });
    })

Cosmos:

• This has through put set to autoScale with 4000 max RUs
• Just add an entry, no upsert is happening on this container.

The main thing I have noticed in logs is, when this container keeps receiving an events, in the time of the delay, I see upto 130 feeds are triggered at once.

I don’t know what is the reason that sometime this container feeds this much events instead of releasing them as soon as it comes.