part: allocate watch channels lazily by joamaki · Pull Request #178 · cilium/statedb

joamaki · 2026-08-21T15:25:32Z

Only allocate the watch channels when requested.

Split Tree and Txn Get/Prefix into no-watch defaults and explicit GetWatch/PrefixWatch methods, and route ordinary StateDB index operations through no-watch paths while preserving existing watch invalidation semantics.

StateDB benchmarks:

goos: linux
goarch: arm64
pkg: github.com/cilium/statedb
                           │    a.txt     │               b.txt                │
                           │    sec/op    │    sec/op     vs base              │
DB_Modify-6                  570.2µ ± ∞ ¹   528.1µ ± ∞ ¹       ~ (p=0.095 n=5)
DB_GetInsert-6               633.9µ ± ∞ ¹   586.5µ ± ∞ ¹       ~ (p=0.095 n=5)
DB_RandomInsert-6            613.6µ ± ∞ ¹   515.3µ ± ∞ ¹       ~ (p=0.095 n=5)
DB_SequentialInsert-6        564.4µ ± ∞ ¹   501.2µ ± ∞ ¹       ~ (p=0.151 n=5)
DB_RandomLookup-6            34.62µ ± ∞ ¹   33.48µ ± ∞ ¹  -3.30% (p=0.016 n=5)
DB_SequentialLookup-6        27.89µ ± ∞ ¹   26.04µ ± ∞ ¹  -6.63% (p=0.008 n=5)
DB_Prefix_SecondaryIndex-6   112.2µ ± ∞ ¹   111.2µ ± ∞ ¹       ~ (p=0.151 n=5)
geomean                      201.7µ         186.2µ        -7.66%
¹ need >= 6 samples for confidence interval at level 0.95

                           │     a.txt     │                 b.txt                  │
                           │     B/op      │     B/op       vs base                 │
DB_Modify-6                  533.5Ki ± ∞ ¹   431.1Ki ± ∞ ¹  -19.18% (p=0.008 n=5)
DB_GetInsert-6               517.8Ki ± ∞ ¹   415.5Ki ± ∞ ¹  -19.76% (p=0.008 n=5)
DB_RandomInsert-6            509.9Ki ± ∞ ¹   407.7Ki ± ∞ ¹  -20.06% (p=0.008 n=5)
DB_SequentialInsert-6        509.9Ki ± ∞ ¹   407.7Ki ± ∞ ¹  -20.06% (p=0.008 n=5)
DB_RandomLookup-6              0.000 ± ∞ ¹     0.000 ± ∞ ¹        ~ (p=1.000 n=5) ²
DB_SequentialLookup-6          0.000 ± ∞ ¹     0.000 ± ∞ ¹        ~ (p=1.000 n=5) ²
DB_Prefix_SecondaryIndex-6   122.0Ki ± ∞ ¹   122.0Ki ± ∞ ¹   +0.03% (p=0.008 n=5)
geomean                                  ³                  -11.82%               ³
¹ need >= 6 samples for confidence interval at level 0.95
² all samples are equal
³ summaries must be >0 to compute geomean

Reconciler benchmark before (best of three runs):

1000000 objects reconciled in 1.35 seconds (batch size 1000) Throughput 740668.25 objects per second
888MB total allocated, 6015187 in-use objects, 338MB bytes in use 2.00user 0.21system 0:01.50elapsed 147%CPU (0avgtext+0avgdata 617884maxresident)k 152inputs+0outputs (1major+203587minor)pagefaults 0swaps

Reconciler benchmark after (best of three runs):

1000000 objects reconciled in 1.32 seconds (batch size 1000) Throughput 760075.67 objects per second
687MB total allocated, 6015138 in-use objects, 239MB bytes in use 2.17user 0.20system 0:01.45elapsed 162%CPU (0avgtext+0avgdata 475600maxresident)k 0inputs+0outputs (0major+136382minor)pagefaults 0swaps

AIL:3

github-actions · 2026-08-21T15:39:38Z

$ make
go build ./...
go: downloading github.com/cilium/hive v1.0.4
go: downloading golang.org/x/time v0.15.0
go: downloading go.yaml.in/yaml/v3 v3.0.4
go: downloading github.com/spf13/cobra v1.10.2
go: downloading github.com/spf13/pflag v1.0.10
go: downloading github.com/cilium/stream v0.0.1
go: downloading github.com/liggitt/tabwriter v0.0.0-20181228230101-89fcab3d43de
go: downloading github.com/spf13/viper v1.18.2
go: downloading go.uber.org/dig v1.17.1
go: downloading golang.org/x/term v0.16.0
go: downloading github.com/davecgh/go-spew v1.1.2-0.20180830191138-d8f796af33cc
go: downloading github.com/mitchellh/mapstructure v1.5.0
go: downloading golang.org/x/sys v0.17.0
go: downloading golang.org/x/tools v0.17.0
go: downloading github.com/spf13/cast v1.6.0
go: downloading github.com/fsnotify/fsnotify v1.7.0
go: downloading github.com/sagikazarmark/slog-shim v0.1.0
go: downloading github.com/spf13/afero v1.11.0
go: downloading github.com/subosito/gotenv v1.6.0
go: downloading github.com/hashicorp/hcl v1.0.0
go: downloading gopkg.in/ini.v1 v1.67.0
go: downloading github.com/magiconair/properties v1.8.7
go: downloading github.com/pelletier/go-toml/v2 v2.1.0
go: downloading gopkg.in/yaml.v3 v3.0.1
go: downloading golang.org/x/text v0.14.0
STATEDB_VALIDATE=1 go test ./... -cover -vet=all -test.count 1
go: downloading github.com/stretchr/testify v1.11.1
go: downloading go.uber.org/goleak v1.3.0
go: downloading golang.org/x/exp v0.0.0-20240119083558-1b970713d09a
go: downloading github.com/pmezard/go-difflib v1.0.1-0.20181226105442-5d4384ee4fb2
ok  	github.com/cilium/statedb	318.891s	coverage: 79.3% of statements
ok  	github.com/cilium/statedb/index	0.004s	coverage: 33.7% of statements
ok  	github.com/cilium/statedb/internal	0.036s	coverage: 42.9% of statements
ok  	github.com/cilium/statedb/lpm	3.056s	coverage: 77.9% of statements
ok  	github.com/cilium/statedb/part	51.727s	coverage: 86.9% of statements
ok  	github.com/cilium/statedb/reconciler	0.270s	coverage: 93.1% of statements
	github.com/cilium/statedb/reconciler/benchmark		coverage: 0.0% of statements
	github.com/cilium/statedb/reconciler/example		coverage: 0.0% of statements
go test -race ./... -test.count 1
ok  	github.com/cilium/statedb	29.815s
ok  	github.com/cilium/statedb/index	1.012s
ok  	github.com/cilium/statedb/internal	1.021s
ok  	github.com/cilium/statedb/lpm	2.146s
ok  	github.com/cilium/statedb/part	28.438s
ok  	github.com/cilium/statedb/reconciler	1.380s
?   	github.com/cilium/statedb/reconciler/benchmark	[no test files]
?   	github.com/cilium/statedb/reconciler/example	[no test files]
go test ./... -bench . -benchmem -test.run xxx
goos: linux
goarch: amd64
pkg: github.com/cilium/statedb
cpu: AMD EPYC 9V74 80-Core Processor                
BenchmarkDB_WriteTxn_1-4                      	  984582	      1147 ns/op	    871472 objects/sec	     704 B/op	      17 allocs/op
BenchmarkDB_WriteTxn_10-4                     	 2654451	       447.8 ns/op	   2233178 objects/sec	     386 B/op	       8 allocs/op
BenchmarkDB_WriteTxn_100-4                    	 3432463	       347.6 ns/op	   2876977 objects/sec	     383 B/op	       7 allocs/op
BenchmarkDB_WriteTxn_1000-4                   	 3097887	       387.1 ns/op	   2583005 objects/sec	     342 B/op	       7 allocs/op
BenchmarkDB_WriteTxn_100_SecondaryIndex-4     	 1404951	       851.5 ns/op	   1174446 objects/sec	     792 B/op	      20 allocs/op
BenchmarkDB_WriteTxn_CommitOnly_100Tables-4   	 1238578	       968.6 ns/op	    1128 B/op	       6 allocs/op
BenchmarkDB_WriteTxn_CommitOnly_1Table-4      	 1912472	       637.0 ns/op	     240 B/op	       6 allocs/op
BenchmarkDB_NewWriteTxn-4                     	 1971141	       592.3 ns/op	     216 B/op	       5 allocs/op
BenchmarkDB_WriteTxnCommit100-4               	 1252765	       957.5 ns/op	    1112 B/op	       6 allocs/op
BenchmarkDB_NewReadTxn-4                      	732564398	         1.638 ns/op	       0 B/op	       0 allocs/op
BenchmarkDB_Modify-4                          	    2924	    405755 ns/op	   2464542 objects/sec	  366869 B/op	    8073 allocs/op
BenchmarkDB_GetInsert-4                       	    2565	    465347 ns/op	   2148933 objects/sec	  350867 B/op	    8073 allocs/op
BenchmarkDB_RandomInsert-4                    	    2882	    394204 ns/op	   2536760 objects/sec	  342863 B/op	    7073 allocs/op
BenchmarkDB_RandomReplace-4                   	     690	   1739141 ns/op	    574997 objects/sec	 1474369 B/op	   28103 allocs/op
BenchmarkDB_SequentialInsert-4                	    3115	    380249 ns/op	   2629853 objects/sec	  342863 B/op	    7073 allocs/op
BenchmarkDB_SequentialInsert_Prefix-4         	     714	   1644679 ns/op	    608021 objects/sec	 3245627 B/op	   45532 allocs/op
BenchmarkDB_Changes_Baseline-4                	    2608	    462360 ns/op	   2162815 objects/sec	  399795 B/op	    9165 allocs/op
BenchmarkDB_Changes-4                         	    1480	    811029 ns/op	   1233001 objects/sec	  600514 B/op	   12319 allocs/op
BenchmarkDB_RandomLookup-4                    	   35648	     33532 ns/op	  29822587 objects/sec	       0 B/op	       0 allocs/op
BenchmarkDB_SequentialLookup-4                	   35659	     33645 ns/op	  29722146 objects/sec	       0 B/op	       0 allocs/op
BenchmarkDB_Prefix_SecondaryIndex-4           	    9644	    114707 ns/op	   8717859 objects/sec	  124952 B/op	    1026 allocs/op
BenchmarkDB_FullIteration_All-4               	    1179	    936264 ns/op	 106807522 objects/sec	     104 B/op	       4 allocs/op
BenchmarkDB_FullIteration_Prefix-4            	    1082	   1011252 ns/op	  98887303 objects/sec	     136 B/op	       5 allocs/op
BenchmarkDB_FullIteration_Get-4               	     285	   4209520 ns/op	  23755677 objects/sec	       0 B/op	       0 allocs/op
BenchmarkDB_FullIteration_Get_Secondary-4     	     139	   8494376 ns/op	  11772495 objects/sec	       0 B/op	       0 allocs/op
BenchmarkDB_FullIteration_ReadTxnGet-4        	     280	   4240523 ns/op	  23581996 objects/sec	       0 B/op	       0 allocs/op
BenchmarkDB_PropagationDelay-4                	  955857	      1179 ns/op	        10.00 50th_µs	        13.00 90th_µs	        32.00 99th_µs	     877 B/op	      20 allocs/op
BenchmarkDB_WriteTxn_100_LPMIndex-4           	  681998	      1737 ns/op	    575635 objects/sec	    1671 B/op	      37 allocs/op
BenchmarkDB_WriteTxn_1_LPMIndex-4             	  173929	     12669 ns/op	     78930 objects/sec	   16010 B/op	      85 allocs/op
BenchmarkDB_LPMIndex_Get-4                    	     472	   2479459 ns/op	   4033139 objects/sec	       0 B/op	       0 allocs/op
BenchmarkWatchSet_4-4                         	 3192320	       370.8 ns/op	     296 B/op	       4 allocs/op
BenchmarkWatchSet_16-4                        	  899860	      1198 ns/op	    1096 B/op	       5 allocs/op
BenchmarkWatchSet_128-4                       	  111554	     10676 ns/op	    8904 B/op	       5 allocs/op
BenchmarkWatchSet_1024-4                      	    9987	    107926 ns/op	   73742 B/op	       5 allocs/op
PASS
ok  	github.com/cilium/statedb	43.655s
PASS
ok  	github.com/cilium/statedb/index	0.003s
goos: linux
goarch: amd64
pkg: github.com/cilium/statedb/internal
cpu: AMD EPYC 9V74 80-Core Processor                
Benchmark_SortableMutex-4   	 7449078	       161.2 ns/op	       0 B/op	       0 allocs/op
PASS
ok  	github.com/cilium/statedb/internal	1.204s
goos: linux
goarch: amd64
pkg: github.com/cilium/statedb/lpm
cpu: AMD EPYC 9V74 80-Core Processor                
Benchmark_txn_insert/batchSize=1-4         	    2556	    457484 ns/op	   2185869 objects/sec	  838425 B/op	   13975 allocs/op
Benchmark_txn_insert/batchSize=10-4        	    4333	    280074 ns/op	   3570490 objects/sec	  385202 B/op	    6668 allocs/op
Benchmark_txn_insert/batchSize=100-4       	    4627	    263176 ns/op	   3799741 objects/sec	  345619 B/op	    6027 allocs/op
Benchmark_txn_delete/batchSize=1-4         	    2041	    583285 ns/op	   1714427 objects/sec	 1286471 B/op	   13976 allocs/op
Benchmark_txn_delete/batchSize=10-4        	    4185	    296657 ns/op	   3370892 objects/sec	  372418 B/op	    5769 allocs/op
Benchmark_txn_delete/batchSize=100-4       	    4689	    254833 ns/op	   3924131 objects/sec	  286753 B/op	    5038 allocs/op
Benchmark_LPM_Lookup-4                     	    9981	    119122 ns/op	   8394780 objects/sec	       0 B/op	       0 allocs/op
Benchmark_LPM_All-4                        	  181165	      6611 ns/op	 151253084 objects/sec	      32 B/op	       1 allocs/op
Benchmark_LPM_Prefix-4                     	  172891	      6921 ns/op	 144484489 objects/sec	      32 B/op	       1 allocs/op
Benchmark_LPM_LowerBound-4                 	  321079	      3660 ns/op	 136613521 objects/sec	     288 B/op	       2 allocs/op
PASS
ok  	github.com/cilium/statedb/lpm	11.996s
goos: linux
goarch: amd64
pkg: github.com/cilium/statedb/part
cpu: AMD EPYC 9V74 80-Core Processor                
Benchmark_Uint64Map_Random-4                  	    2182	    553296 ns/op	   1807351 items/sec	 2424065 B/op	    6043 allocs/op
Benchmark_Uint64Map_Sequential-4              	    2610	    455972 ns/op	   2193119 items/sec	 2112690 B/op	    5753 allocs/op
Benchmark_Uint64Map_Sequential_Insert-4       	    2829	    420017 ns/op	   2380856 items/sec	 2104686 B/op	    4752 allocs/op
Benchmark_Uint64Map_Sequential_Txn_Insert-4   	   14948	     78653 ns/op	  12714023 items/sec	   86112 B/op	    2027 allocs/op
Benchmark_Uint64Map_Random_Insert-4           	    2188	    509897 ns/op	   1961179 items/sec	 2412209 B/op	    5030 allocs/op
Benchmark_Uint64Map_Random_Txn_Insert-4       	    9648	    123945 ns/op	   8068096 items/sec	  119964 B/op	    2420 allocs/op
Benchmark_Insert_RootOnlyWatch-4              	   15308	     78208 ns/op	  12786433 objects/sec	   71264 B/op	    2032 allocs/op
Benchmark_Insert-4                            	   12834	     93434 ns/op	  10702715 objects/sec	   79680 B/op	    3058 allocs/op
Benchmark_WatchReplace-4                      	   14048	     85508 ns/op	  11694872 objects/sec	   65345 B/op	    2011 allocs/op
Benchmark_Modify-4                            	   17642	     67968 ns/op	  14712764 objects/sec	   58120 B/op	    1007 allocs/op
Benchmark_GetInsert-4                         	   13785	     86969 ns/op	  11498301 objects/sec	   58120 B/op	    1007 allocs/op
Benchmark_Replace-4                           	44561740	        26.84 ns/op	  37261408 objects/sec	       0 B/op	       0 allocs/op
Benchmark_Replace_RootOnlyWatch-4             	44610747	        27.10 ns/op	  36898317 objects/sec	       0 B/op	       0 allocs/op
Benchmark_txn_1-4                             	 8928412	       133.6 ns/op	   7485641 objects/sec	      64 B/op	       3 allocs/op
Benchmark_txn_10-4                            	14962221	        79.57 ns/op	  12567865 objects/sec	      76 B/op	       2 allocs/op
Benchmark_txn_100-4                           	16871628	        70.18 ns/op	  14249830 objects/sec	      79 B/op	       2 allocs/op
Benchmark_txn_1000-4                          	14671488	        79.70 ns/op	  12547768 objects/sec	      65 B/op	       2 allocs/op
Benchmark_txn_delete_1-4                      	 7577736	       158.9 ns/op	   6291317 objects/sec	     632 B/op	       3 allocs/op
Benchmark_txn_delete_10-4                     	16349349	        72.51 ns/op	  13790563 objects/sec	     103 B/op	       1 allocs/op
Benchmark_txn_delete_100-4                    	17038371	        64.22 ns/op	  15571888 objects/sec	      47 B/op	       1 allocs/op
Benchmark_txn_delete_1000-4                   	20348007	        58.44 ns/op	  17112450 objects/sec	      24 B/op	       1 allocs/op
Benchmark_Get-4                               	   59886	     20164 ns/op	  49593639 objects/sec	       0 B/op	       0 allocs/op
Benchmark_GetWatch-4                          	   57576	     20859 ns/op	  47939952 objects/sec	       0 B/op	       0 allocs/op
Benchmark_All-4                               	  227210	      5550 ns/op	 180166195 objects/sec	       0 B/op	       0 allocs/op
Benchmark_Iterator_All-4                      	  205575	      5832 ns/op	 171481742 objects/sec	       0 B/op	       0 allocs/op
Benchmark_Iterator_Next-4                     	  196029	      6055 ns/op	 165152692 objects/sec	     896 B/op	       1 allocs/op
Benchmark_Hashmap_Insert-4                    	   22927	     52449 ns/op	  19066160 objects/sec	   74264 B/op	      20 allocs/op
Benchmark_Hashmap_Get_Uint64-4                	  198238	      6107 ns/op	 163740408 objects/sec	       0 B/op	       0 allocs/op
Benchmark_Hashmap_Get_Bytes-4                 	  162464	      7393 ns/op	 135256801 objects/sec	       0 B/op	       0 allocs/op
Benchmark_Delete_Random-4                     	     100	  10656220 ns/op	   9384190 objects/sec	 2111812 B/op	  102363 allocs/op
Benchmark_find16-4                            	292483246	         4.099 ns/op	       0 B/op	       0 allocs/op
Benchmark_findIndex16-4                       	100000000	        11.18 ns/op	       0 B/op	       0 allocs/op
Benchmark_find48-4                            	485319165	         2.461 ns/op	       0 B/op	       0 allocs/op
Benchmark_findIndex48_hit-4                   	549089155	         2.084 ns/op	       0 B/op	       0 allocs/op
Benchmark_findIndex48_miss-4                  	487581696	         2.461 ns/op	       0 B/op	       0 allocs/op
Benchmark_find4-4                             	545672948	         2.198 ns/op	       0 B/op	       0 allocs/op
Benchmark_findIndex4-4                        	399614042	         3.004 ns/op	       0 B/op	       0 allocs/op
PASS
ok  	github.com/cilium/statedb/part	44.810s
PASS
ok  	github.com/cilium/statedb/reconciler	0.004s
?   	github.com/cilium/statedb/reconciler/benchmark	[no test files]
?   	github.com/cilium/statedb/reconciler/example	[no test files]
go run ./reconciler/benchmark -quiet
1000000 objects reconciled in 1.44 seconds (batch size 1000)
Throughput 693928.90 objects per second
616MB total allocated, 6015096 in-use objects, 239MB bytes in use

Only allocate the watch channels when requested. Split Tree and Txn Get/Prefix into no-watch defaults and explicit GetWatch/PrefixWatch methods, and route ordinary StateDB index operations through no-watch paths while preserving existing watch invalidation semantics. StateDB benchmarks: goos: linux goarch: arm64 pkg: github.com/cilium/statedb │ a.txt │ b.txt │ │ sec/op │ sec/op vs base │ DB_Modify-6 570.2µ ± ∞ ¹ 528.1µ ± ∞ ¹ ~ (p=0.095 n=5) DB_GetInsert-6 633.9µ ± ∞ ¹ 586.5µ ± ∞ ¹ ~ (p=0.095 n=5) DB_RandomInsert-6 613.6µ ± ∞ ¹ 515.3µ ± ∞ ¹ ~ (p=0.095 n=5) DB_SequentialInsert-6 564.4µ ± ∞ ¹ 501.2µ ± ∞ ¹ ~ (p=0.151 n=5) DB_RandomLookup-6 34.62µ ± ∞ ¹ 33.48µ ± ∞ ¹ -3.30% (p=0.016 n=5) DB_SequentialLookup-6 27.89µ ± ∞ ¹ 26.04µ ± ∞ ¹ -6.63% (p=0.008 n=5) DB_Prefix_SecondaryIndex-6 112.2µ ± ∞ ¹ 111.2µ ± ∞ ¹ ~ (p=0.151 n=5) geomean 201.7µ 186.2µ -7.66% ¹ need >= 6 samples for confidence interval at level 0.95 │ a.txt │ b.txt │ │ B/op │ B/op vs base │ DB_Modify-6 533.5Ki ± ∞ ¹ 431.1Ki ± ∞ ¹ -19.18% (p=0.008 n=5) DB_GetInsert-6 517.8Ki ± ∞ ¹ 415.5Ki ± ∞ ¹ -19.76% (p=0.008 n=5) DB_RandomInsert-6 509.9Ki ± ∞ ¹ 407.7Ki ± ∞ ¹ -20.06% (p=0.008 n=5) DB_SequentialInsert-6 509.9Ki ± ∞ ¹ 407.7Ki ± ∞ ¹ -20.06% (p=0.008 n=5) DB_RandomLookup-6 0.000 ± ∞ ¹ 0.000 ± ∞ ¹ ~ (p=1.000 n=5) ² DB_SequentialLookup-6 0.000 ± ∞ ¹ 0.000 ± ∞ ¹ ~ (p=1.000 n=5) ² DB_Prefix_SecondaryIndex-6 122.0Ki ± ∞ ¹ 122.0Ki ± ∞ ¹ +0.03% (p=0.008 n=5) geomean ³ -11.82% ³ ¹ need >= 6 samples for confidence interval at level 0.95 ² all samples are equal ³ summaries must be >0 to compute geomean Reconciler benchmark before (best of three runs): 1000000 objects reconciled in 1.35 seconds (batch size 1000) Throughput 740668.25 objects per second 888MB total allocated, 6015187 in-use objects, 338MB bytes in use 2.00user 0.21system 0:01.50elapsed 147%CPU (0avgtext+0avgdata 617884maxresident)k 152inputs+0outputs (1major+203587minor)pagefaults 0swaps Reconciler benchmark after (best of three runs): 1000000 objects reconciled in 1.32 seconds (batch size 1000) Throughput 760075.67 objects per second 687MB total allocated, 6015138 in-use objects, 239MB bytes in use 2.17user 0.20system 0:01.45elapsed 162%CPU (0avgtext+0avgdata 475600maxresident)k 0inputs+0outputs (0major+136382minor)pagefaults 0swaps AIL:3 Signed-off-by: Jussi Maki <jussi@isovalent.com>

Reconciliation status is committed before metrics are published, and refresh jobs register their health nodes independently. Wait for both states in the multi-reconciler test so it does not depend on goroutine scheduling under the race detector. AIL:3 Signed-off-by: Jussi Maki <jussi@isovalent.com>

Add a benchmark that replaces 1,000 values per committed transaction in a tree with per-node watches. This exercises path cloning and watch notification collection. AIL:3 Signed-off-by: Jussi Maki <jussi@isovalent.com>

Collect watch states in a slice instead of a map. Tree traversal visits each cloned node once. Lazy watches may share a state across physical nodes; closing a state is idempotent, so duplicates are safe. Clear the slice before reuse so closed channels and their subtrees are not retained. goos: linux goarch: arm64 pkg: github.com/cilium/statedb/part │ before.txt │ after.txt │ │ sec/op │ sec/op vs base │ _Insert-6 90.81µ ± 5% 87.97µ ± 2% ~ (p=0.315 n=10) PatchWatchReplace-6 142.30µ ± 8% 81.95µ ± 3% -42.41% (p=0.000 n=10) │ before.txt │ after.txt │ │ objects/sec │ objects/sec vs base │ _Insert-6 11.01M ± 5% 11.37M ± 2% ~ (p=0.315 n=10) PatchWatchReplace-6 7.030M ± 7% 12.202M ± 3% +73.58% (p=0.000 n=10) │ before.txt │ after.txt │ │ B/op │ B/op vs base │ _Insert-6 78.66Ki ± 0% 77.81Ki ± 0% -1.07% (p=0.000 n=10) PatchWatchReplace-6 136.52Ki ± 0% 63.81Ki ± 0% -53.26% (p=0.000 n=10) │ before.txt │ after.txt │ │ allocs/op │ allocs/op vs base │ _Insert-6 3.060k ± 0% 3.058k ± 0% -0.07% (p=0.000 n=10) PatchWatchReplace-6 2.033k ± 0% 2.011k ± 0% -1.08% (p=0.000 n=10) AIL:3 Signed-off-by: Jussi Maki <jussi@isovalent.com>

pippolo84

✔️

joamaki requested a review from a team as a code owner August 21, 2026 15:25

joamaki requested review from derailed and removed request for a team August 21, 2026 15:25

joamaki mentioned this pull request Aug 24, 2026

Small performance improvements #181

Merged

joamaki added 4 commits August 25, 2026 11:59

part: benchmark watch replacement transactions

dacff79

Add a benchmark that replaces 1,000 values per committed transaction in a tree with per-node watches. This exercises path cloning and watch notification collection. AIL:3 Signed-off-by: Jussi Maki <jussi@isovalent.com>

joamaki force-pushed the pr/joamaki/lazy-watches branch from 1084d0c to 8c26bd9 Compare August 25, 2026 09:59

pippolo84 approved these changes Aug 25, 2026

View reviewed changes

pippolo84 removed the request for review from derailed August 25, 2026 10:45

joamaki merged commit b597c01 into main Aug 25, 2026
1 check passed

joamaki deleted the pr/joamaki/lazy-watches branch August 25, 2026 11:33

Navigation Menu

Uh oh!

part: allocate watch channels lazily - #178

part: allocate watch channels lazily#178
joamaki merged 4 commits into
mainfrom
pr/joamaki/lazy-watches

joamaki commented Aug 21, 2026

Uh oh!

github-actions Bot commented Aug 21, 2026 •

edited

Loading

Uh oh!

pippolo84 left a comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

joamaki commented Aug 21, 2026

Uh oh!

github-actions Bot commented Aug 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pippolo84 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

github-actions Bot commented Aug 21, 2026 •

edited

Loading