==== next.orly.dev Profiling Analysis ==== Sat Sep 13 11:46:40 UTC 2025 -- CPU Hotspots (top by flat CPU) -- Type: cpu Time: Sep 13, 2025 at 11:39am (UTC) Duration: 400.39s, Total samples = 2620ms ( 0.65%) Showing nodes accounting for 1950ms, 74.43% of 2620ms total Dropped 58 nodes (cum <= 13.10ms) Showing top 15 nodes out of 78 flat flat% sum% cum cum% 870ms 33.21% 33.21% 870ms 33.21% internal/runtime/syscall.Syscall6 220ms 8.40% 41.60% 220ms 8.40% runtime.futex 210ms 8.02% 49.62% 400ms 15.27% runtime.stealWork 160ms 6.11% 55.73% 160ms 6.11% runtime.write1 70ms 2.67% 58.40% 70ms 2.67% runtime.nanotime1 60ms 2.29% 60.69% 60ms 2.29% runtime.pMask.read (inline) 50ms 1.91% 62.60% 380ms 14.50% github.com/dgraph-io/badger/v4.(*levelsController).runCompactor 50ms 1.91% 64.50% 50ms 1.91% runtime.(*randomEnum).next (inline) 50ms 1.91% 66.41% 1980ms 75.57% runtime.findRunnable 40ms 1.53% 67.94% 40ms 1.53% runtime.memclrNoHeapPointers 40ms 1.53% 69.47% 930ms 35.50% runtime.netpoll 40ms 1.53% 70.99% 150ms 5.73% runtime.selectgo 30ms 1.15% 72.14% 860ms 32.82% internal/runtime/syscall.EpollWait 30ms 1.15% 73.28% 40ms 1.53% runtime.(*mcache).prepareForSweep 30ms 1.15% 74.43% 30ms 1.15% runtime.cheaprand (inline) -- Memory (In-Use Space) Hotspots -- Type: inuse_space Time: Sep 13, 2025 at 11:46am (UTC) Showing nodes accounting for 85209.25kB, 99.24% of 85865.58kB total Dropped 112 nodes (cum <= 429.33kB) flat flat% sum% cum cum% 85200kB 99.22% 99.22% 85200kB 99.22% github.com/dgraph-io/badger/v4/skl.newArena (inline) 9.25kB 0.011% 99.24% 85654.66kB 99.75% github.com/dgraph-io/badger/v4.Open 0 0% 99.24% 85200kB 99.22% github.com/dgraph-io/badger/v4.(*DB).newMemTable 0 0% 99.24% 85200kB 99.22% github.com/dgraph-io/badger/v4.(*DB).openMemTable 0 0% 99.24% 85200kB 99.22% github.com/dgraph-io/badger/v4/skl.NewSkiplist 0 0% 99.24% 85654.66kB 99.75% main.main 0 0% 99.24% 85654.66kB 99.75% next.orly.dev/pkg/database.New 0 0% 99.24% 85654.66kB 99.75% runtime.main -- Allocations (Total Alloc Space) Hotspots -- Type: alloc_space Time: Sep 13, 2025 at 11:46am (UTC) Showing nodes accounting for 83.20MB, 100% of 83.21MB total Dropped 26 nodes (cum <= 0.42MB) flat flat% sum% cum cum% 83.20MB 100% 100% 83.20MB 100% github.com/dgraph-io/badger/v4/skl.newArena (inline) 0 0% 100% 83.20MB 100% github.com/dgraph-io/badger/v4.(*DB).newMemTable 0 0% 100% 83.20MB 100% github.com/dgraph-io/badger/v4.(*DB).openMemTable 0 0% 100% 83.20MB 100% github.com/dgraph-io/badger/v4.Open 0 0% 100% 83.20MB 100% github.com/dgraph-io/badger/v4/skl.NewSkiplist 0 0% 100% 83.20MB 100% main.main 0 0% 100% 83.20MB 100% next.orly.dev/pkg/database.New 0 0% 100% 83.20MB 100% runtime.main -- Allocation Frequency (Alloc Objects) -- Type: alloc_objects Time: Sep 13, 2025 at 11:46am (UTC) Showing nodes accounting for 7, 100% of 7 total Showing top 15 nodes out of 34 flat flat% sum% cum cum% 2 28.57% 28.57% 2 28.57% regexp/syntax.(*parser).newRegexp (inline) 1 14.29% 42.86% 1 14.29% github.com/dgraph-io/badger/v4.(*levelsController).levelTargets 1 14.29% 57.14% 1 14.29% github.com/dgraph-io/badger/v4/skl.newArena (inline) 1 14.29% 71.43% 1 14.29% github.com/pkg/profile.Start 1 14.29% 85.71% 1 14.29% github.com/pkg/profile.Start.func11 1 14.29% 100% 1 14.29% runtime.allocm 0 0% 100% 1 14.29% github.com/dgraph-io/badger/v4.(*DB).newMemTable 0 0% 100% 1 14.29% github.com/dgraph-io/badger/v4.(*DB).openMemTable 0 0% 100% 1 14.29% github.com/dgraph-io/badger/v4.(*levelsController).pickCompactLevels 0 0% 100% 1 14.29% github.com/dgraph-io/badger/v4.(*levelsController).runCompactor 0 0% 100% 1 14.29% github.com/dgraph-io/badger/v4.(*levelsController).runCompactor.func3 0 0% 100% 1 14.29% github.com/dgraph-io/badger/v4.Open 0 0% 100% 1 14.29% github.com/dgraph-io/badger/v4/skl.NewSkiplist 0 0% 100% 2 28.57% github.com/google/pprof/profile.init 0 0% 100% 2 28.57% main.main Notes: - CPU section identifies functions using the most CPU time. - Memory section identifies which functions retain the most memory (in-use). - Allocations sections identify functions responsible for the most allocation volume and count, which correlates with GC pressure. - Profiles are created by github.com/pkg/profile and may only be flushed when the relay process receives a shutdown; CPU profile often requires process exit.