Provenance queries effect on processing

2020-11-03 Thread Eric Secules
Hello everyone, I was wondering if it were possible for excessive use of the provenance api would cause flowfile processing to slow down and even come to a halt? My test setup queries the provenance API to see if all flowfiles that descended from a given input file have completed processing. This

Re: Provenance queries effect on processing

2020-11-03 Thread Joe Witt
Eric, short version: Provenance queries can absolutely take away CPU time from the flow. longer version: They get the same priority as any other thread in nifi. Once prov queries are being executed they use CPU. I would strongly advise against any blending of the flow execution with provenance

Re: Provenance queries effect on processing

2020-11-03 Thread Eric Secules
Hi Joe, Thanks for the explanation, is there a Jira ticket for for a job management mechanism? Is this a priority for a coming release? Is there a lag between events occurring and them becoming searchable, if so what settings help control this lag? Thanks, Eric On Tue, Nov 3, 2020 at 2:35 PM Jo

Re: Provenance queries effect on processing

2020-11-04 Thread Eric Secules
Hello, I agree it's not the best idea to use the provenance data constantly to check when a test file is finally done being processed. Do you know if anything came out in 1.12.0 or 1.12.1 which would cause provenance queries to have a greater impact? We recently upgraded nifi among other changes a

Re: Provenance queries effect on processing

2020-11-04 Thread Joe Witt
Eric Nope Im not aware of anything specific. Thanks Joe On Wed, Nov 4, 2020 at 9:54 AM Eric Secules wrote: > Hello, > > I agree it's not the best idea to use the provenance data constantly to > check when a test file is finally done being processed. Do you know if > anything came out in 1.12.0

Re: Provenance queries effect on processing

2020-11-04 Thread Bryan Bende
Not sure if this would be helpful for you, but I wonder if you could use the new process group concurrency controls introduced in 1.12.0 to help determine when processing is complete. You could set a process group to allow 1 flow file in at a time and then batch output to wait for all flow files to

Re: Provenance queries effect on processing

2020-11-16 Thread Eric Secules
I dont think this solution would work for me. I am wondering whether the rollover time has any affect on when I am able to query the events. Does the provenance log file need to roll over before it's indexed? On Wed, Nov 4, 2020 at 11:21 AM Bryan Bende wrote: > Not sure if this would be helpful