您正在创建太多 HashedWheelTimer 实例

Question

I am using java sdk:

compile(group: 'com.azure',  name: 'azure-cosmos', version: '4.8.0')

in azure functions. Don't know from where the above error is coming from:

You are creating too many HashedWheelTimer instances. HashedWheelTimer is a shared resource that must be reused across the JVM,so that only a few instances are created.

Answer 1

Look, that's not full answer as you expect - this is try to help and maybe it will give you any ideas how to go further with this very uncommon error..

I think an error go from this part: reporting ReportTooManyInstances https://github.com/Azure/DotNetty/blob/4a76e624580e0005257513fce1c69555294763f5/src/DotNetty.Common/Utilities/HashedWheelTimer.cs#L257-L258

  static void ReportTooManyInstances() =>
        Logger.Error($"You are creating too many {nameof(HashedWheelTimer)} instances. 
  {nameof(HashedWheelTimer)} is a shared resource that must be reused across the process,so that only a few instances are created.");

There is a hardcode for InstanceCountLimit = 64 in the begging: https://github.com/Azure/DotNetty/blob/4a76e624580e0005257513fce1c69555294763f5/src/DotNetty.Common/Utilities/HashedWheelTimer.cs#L19-L27

    public sealed class HashedWheelTimer : ITimer
    {
        static readonly IInternalLogger Logger =
            InternalLoggerFactory.GetInstance<HashedWheelTimer>();

        static int instanceCounter;
        static int warnedTooManyInstances;

        const int InstanceCountLimit = 64;

And if you reach InstanceCountLimit , there is a piece of code stating the ReportTooManyInstances should be thrown.

https://github.com/Azure/DotNetty/blob/4a76e624580e0005257513fce1c69555294763f5/src/DotNetty.Common/Utilities/HashedWheelTimer.cs#L105-L108

            if (Interlocked.Increment(ref instanceCounter) > InstanceCountLimit &&
                Interlocked.CompareExchange(ref warnedTooManyInstances, 1, 0) == 0)
            {
                ReportTooManyInstances();
            }

What I also found(maybe will be interesting for you):

Functions best practices:Optimize the performance and reliability of Azure Functions