You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
As mentioned in the wikipedia on cronspecs, Jenkins supports a new field value 'H' (or 'H/15', etc.) which is used to specify a random stable offset within the schedule. Jenkins hashes (hence the 'H') the job name to to allow you to give the same schedule to multiple jobs without running into the thundering herd problem.
In order to support this, the API would need to change to allow passing in a job name to use for hashing.
This would make it easier to schedule recurring jobs on kubernetes.
The text was updated successfully, but these errors were encountered:
Any updates to this issue? Top-of-the-hour thundering herd is an easily preventable problem. I would love to change @hourly to H * * * * * where H renders to the hash of the Job Name, PID, hostname, or something pseudo-random. Or, instead of H, R for random.
As mentioned in the wikipedia on cronspecs, Jenkins supports a new field value 'H' (or 'H/15', etc.) which is used to specify a random stable offset within the schedule. Jenkins hashes (hence the 'H') the job name to to allow you to give the same schedule to multiple jobs without running into the thundering herd problem.
In order to support this, the API would need to change to allow passing in a job name to use for hashing.
This would make it easier to schedule recurring jobs on kubernetes.
The text was updated successfully, but these errors were encountered: