Needs better dampening?

[markpundsack]

With no traffic on an app running on my local box, it keeps adding, then removing, then adding workers.

15:03:48 web.1    | PumaAutoTune (105s): Potential memory leak. Reaping largest worker puma.largest_worker_memory_mb=106.13671875 puma.resource_ram_mb=424.1015625 puma.current_cluster_size=4
15:03:48 web.1    | PumaAutoTune (105s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=429.6328125 puma.current_cluster_size=4
15:03:49 web.1    | PumaAutoTune (105s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=450.79296875 puma.current_cluster_size=4
15:03:49 web.1    | PumaAutoTune (106s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=341.35546875 puma.current_cluster_size=4
15:03:59 web.1    | PumaAutoTune (116s): Cluster too small. Resizing to add one more worker puma.resource_ram_mb=228.1171875 puma.current_cluster_size=2
15:04:09 web.1    | PumaAutoTune (126s): Cluster too small. Resizing to add one more worker puma.resource_ram_mb=228.19921875 puma.current_cluster_size=2
15:04:12 web.1    | [74805] + Gemfile in context: /Users/mpundsack/Dropbox/Heroku/src/dashboard/Gemfile
15:04:12 web.1    | [72445] - Worker 0 (pid: 74805) booted, phase: 0
15:04:19 web.1    | PumaAutoTune (136s): Potential memory leak. Reaping largest worker puma.largest_worker_memory_mb=96.01171875 puma.resource_ram_mb=317.01171875 puma.current_cluster_size=3
15:04:19 web.1    | PumaAutoTune (136s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=321.58984375 puma.current_cluster_size=3
15:04:19 web.1    | PumaAutoTune (136s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=339.21875 puma.current_cluster_size=3
15:04:19 web.1    | PumaAutoTune (136s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=237.6640625 puma.current_cluster_size=3
15:04:30 web.1    | PumaAutoTune (146s): Cluster too small. Resizing to add one more worker puma.resource_ram_mb=132.56640625 puma.current_cluster_size=1
15:04:32 web.1    | [74923] + Gemfile in context: /Users/mpundsack/Dropbox/Heroku/src/dashboard/Gemfile
15:04:32 web.1    | [72445] - Worker 0 (pid: 74923) booted, phase: 0
15:04:40 web.1    | PumaAutoTune (157s): Cluster too small. Resizing to add one more worker puma.resource_ram_mb=221.703125 puma.current_cluster_size=2
15:04:42 web.1    | [74947] + Gemfile in context: /Users/mpundsack/Dropbox/Heroku/src/dashboard/Gemfile
15:04:42 web.1    | [72445] - Worker 1 (pid: 74947) booted, phase: 0
15:04:50 web.1    | PumaAutoTune (167s): Potential memory leak. Reaping largest worker puma.largest_worker_memory_mb=89.16796875 puma.resource_ram_mb=310.953125 puma.current_cluster_size=3
15:04:50 web.1    | PumaAutoTune (167s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=315.99609375 puma.current_cluster_size=3
15:04:50 web.1    | PumaAutoTune (167s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=329.609375 puma.current_cluster_size=3
15:04:50 web.1    | PumaAutoTune (167s): Cluster too large. Resizing to remove one worker puma.resource_ram_mb=221.98828125 puma.current_cluster_size=3
15:05:00 web.1    | PumaAutoTune (177s): Cluster too small. Resizing to add one more worker puma.resource_ram_mb=222.0625 puma.current_cluster_size=2
15:05:10 web.1    | PumaAutoTune (187s): Cluster too small. Resizing to add one more worker puma.resource_ram_mb=222.1796875 puma.current_cluster_size=2

[markpundsack]

Also, why do I get multiple messages with the same cluster target size?