We are looking into similar metrics. Did you find anything out? We use Filebeat to forward and centralize logs to ELK. With that data, we have a few queries to find passed and failed jobs. However, I cannot find any log entries that can help us answer how long the queue for agents is. https://gocd-server/go/api/jobs/scheduled.xml will give an XML. I also saw that there's a performance log entry in log4j.properties, but I couldn't see how that would help.
On Monday, April 18, 2016 at 6:22:45 PM UTC+2, [email protected] wrote: > > Hello, > > I'm conducting a deep analysis of our pipeline performance which has drawn > focus to the various stages of a Go stage/job execution. E.g how long a job > takes to schedule to an agent, vs. 'preparing' vs. building etc > > I'd like to start maintaining these metrics on every stage of every > pipeline without hitting the API hundreds of times a second. > > Does anyone have any advice for instrumenting Go like this? > > A few ideas talked about so far: > - Explore the option of performance logging (if this exists) and pump > through to ELK, graph with Kibana > - Dump the Postgres DB to a backup and graph the results of multiple > queries > > Thanks, > Max > -- You received this message because you are subscribed to the Google Groups "go-cd" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
