Currently, we have a few tasks that may fail on their initial run, but are highly likely to succeed if we retry them once or twice. As a result, we are using the "On Error, Wait x seconds and Retry Task" option within the task details. However, we get e-mail notifications saying the task failed (which I guess technically it did the first time) even though it succeeded on the second or third run. We are using the "Notify on Error" notification at the Job level, not at the Task level. Do you know of a work around so that we will only be notified if the last re-try of the task fails? In other words, we don't want a notification if the task fails, fails, and then succeeds, but we would want a notification if the task fails, fails, fails (with two retry's for example).

I am wondering if there is a way to accomplish something similar if we use conditions and repeat the task, assuming there is not a way to accomplish what we desire using the wait..retry task option.

Any help would be greatly appreciated since we currently receive production alerts that we have to investigate only to find out that they succeeded on the second or third try.

Thank you.

Yes, the only way is to have the complete Notification on the Task - and you will get these error emails before - as you currently have it right now. To implement this would require a lot of internal changes and I am not sure if it is worth it in this case.

---

Thanks for the feedback. Some additional details I should have mentioned: First, for our production jobs, when a job fails, we send an e-mail to our on-call pagers, so we wind up waking our support staff up in the middle of the night to investigate, but only to find out that the job eventually succeeded. Second, we have at least 5 production jobs that can fail in this way and they run every 5 minutes, so adding a completion e-mail will be pretty cumbersome to get 60 e-mails an hour.

Would we be able to create a work around by using something other than the notification on error to send us an e-mail? This is where I don't know much about conditions and some other features of VisualCron, so I'm looking for some help. For example, could we have a job with no notifications, but starts with Task #1 that runs command line A. Then, if Task #1 fails, we have a task right after it (Task #2) that runs command line A. Then if Task #2 fails, we have another task (Task #3) that also runs command line A. Then, if Task #3 fails, we have another task (Task #4) that is a type=email and that task (#4) sends us an e-mail to our on-call pager and will wake up our on-call person in the middle of the night (because command line A failed three times). Also, if Task #1, #2, or #3 succeeds, we would need to stop executing and ensure we skip e-mail task #4. Are we able to do something similar to what I describe here? Another variation is that we could move Task #4 (e-mail our pager) to a different Job altogether if that would make it easier.

Thank you.

I think in your case, the simpliest approach is to have an Error Notification on the Job but not the Task. So, it will retry a number of times but if the Task finally fails you will get the Error Notification from the Job instead.

In the Error Notification you can use Variables in the message that shows last runtime, task name and last result so you can easily determine which Task failed.

Will that solve it for you?

---

Henrik,

That would work perfectly and that is how we have it set up (the Error Notification is attached to the Job, not the task). However, when the task fails, fails, succeeds (on auto-retry), the error notification on the Job still fires and tells us that the Job failed. Is that something you could change? Or is there a way we can configure it to behave that way? Would adding another dummy task after it that always succeeds help?

Thank you.

We will check that. It should be a success if it is a retry that finally succeeds - in my opinion. Will get back to you.

---

It seems this is the exact problem we are having ... a task fails on first attempt, but then succeeds on a retry, and everything completes correctly, but at the job level, the "failure" email notification still fires off.

I see this thread is over a year old ... has this issue ever been resolved, or is there now some way in VC to modify the "failure" notification behavior?

Thanks

This issue is fixed. There is no longer an error at Job level if it fails and then later succeeds.

---

What release version is this corrected in?

I am not sure which version this was fixed in. Somehow we have missed to document that.

---