The agent’s workers limit has reached

General Questions
1

Hi Team,

We have a couple of Jobs scheduled at non-overlapping time windows.
Recently we have upgraded our agent to the latest 3.1 and now we are getting the following issue when trying to run our job.

"The job failed to start due to the following reason:
The agent’s workers limit has reached. The execution is queued"

Once this issue occurs the pending jobs fail to execute.
We are not able to find any help on any forum regarding this issue…Can you please help ?

1 Like
2

FYI found a way to restrict the max-number of workers. This can be done using CLI, just pass the parameter --max-workers with the start command.

E.g ./testproject-agent start --max-workers 1

1 Like
3

getting the same problem - any solution?

4

Hi,
When agent reports job was queued because max workers has been reached, can you see it queued on the agents tab here?

image
image1920×937 142 KB

And if you can see the queued job do they eventually start when the agent is idle?

5

Same issue here, and no they just sit in the queue when the agent is idle until you manually delete them. Is there a way to not have exceptions added to the queue?

6

Hi,
When the agent busy all jobs / test execution will be queued. we have an issue with queued jobs currently more info here:

I will reply here once the issue is resolved.

7

Hi,
The issue has been identified and resolved. 201 is now documented in our swagger at:
https://api.testproject.io/docs/v2/#/

Also queued jobs should immediately start when agent workers are available.

Let me know if there is any issue.

8

@Amit.Lacher:

It seems like the issue still persists. I’m on v3.3.0, and facing the same issues.

9

A similar thing happens to us from time to time. A job running on the remote agent just starts throwing an error in the middle of the run. There is no error description or anything. When the job is ‘done’, the agent is still marked as busy and it won’t run the next scheduled job. It will just queue the jobs. The agent has to be manually restarted to get to the idle state. Using version 3.3.0 (but happened before), running headless Chrome on Windows. Any advice? Thanks

10

@aleksandar @parag.gupta
Please contact us directly on our live chat which is found in the right bottom corner.
Included in your message:

  • Agent logs
  • Link to this thread
  • Description of the issue and timestamp that pinpoints the issue in the logs.
11

Same here. Really annoying

12

I too have been recently experiencing the same agent issue:
The job failed to start due to the following reason:
The agent’s workers limit has reached. The execution is queued

13

Hi, @rhernandez @r_filmustage,
If your agent reached max workers the message stated is correct and it’s not an error.
If the agent is busy working on a different execution or the worker is still finalizing or uploading the report of the latest execution, the message Job failed to start due to “The agent’s workers limit has reached. The execution is queued” is completely by design.

If you are experiencing a different issue where the agent shows 0/1 workers on the UI but the execution still does not start, or you are experiencing an issue when execution is getting stuck and the agent keeps reporting 1/1. (has been resolved)
Please contact us on any of our support channels.
These are different issues and are not related to the message which correctly states that the execution was queued.
If you believe you get the message wrongly without a proper reason, please contact us with the agent logs and screenshots that will help us get a better understanding of the issue.

Amit.

1 Like
14

Thanks for clarification @Amit.Lacher
Is it possible to set max workers via ENV_VAR for docker instance?

15

Yes, it is possible, see more here:

You will be able to run firefox and chrome at the same time.