Skip to content

Releases: zalando/skipper

v0.21.209

26 Sep 09:42
08a295f
Compare
Choose a tag to compare

Changes

build(deps): bump alpine from 7eccd4d to beefdbd in /packaging (#3250)

Bumps alpine from 7eccd4d to beefdbd.


updated-dependencies:

  • dependency-name: alpine
    dependency-type: direct:production
    ...

Multiarch Docker image

Multiarch Docker image is available in Github's docker registry:

docker run -it ghcr.io/zalando/skipper:v0.21.209 skipper --help

Docker image

Docker image is available in Zalando's Open Source registry:

docker run -it registry.opensource.zalan.do/teapot/skipper:v0.21.209 skipper --help

v0.21.208

25 Sep 08:13
2e9e3cf
Compare
Choose a tag to compare

Changes

minor: remove unused endpointRegistry field from healthyEndpoints and fadeIn (#3248)

Multiarch Docker image

Multiarch Docker image is available in Github's docker registry:

docker run -it ghcr.io/zalando/skipper:v0.21.208 skipper --help

Docker image

Docker image is available in Zalando's Open Source registry:

docker run -it registry.opensource.zalan.do/teapot/skipper:v0.21.208 skipper --help

v0.21.207

25 Sep 07:26
1db1ac7
Compare
Choose a tag to compare

Changes

build(deps): bump alpine from beefdbd to 7eccd4d in /packaging (#3249)

Multiarch Docker image

Multiarch Docker image is available in Github's docker registry:

docker run -it ghcr.io/zalando/skipper:v0.21.207 skipper --help

Docker image

Docker image is available in Zalando's Open Source registry:

docker run -it registry.opensource.zalan.do/teapot/skipper:v0.21.207 skipper --help

v0.21.206

24 Sep 17:17
8ea6c5b
Compare
Choose a tag to compare

Changes

docs: document single argument variant of tracingBaggageToTag (#3247)

Implementation allows single argument variant which
is used in the wild so document it.

Follow up on #1160 and #1168

Multiarch Docker image

Multiarch Docker image is available in Github's docker registry:

docker run -it ghcr.io/zalando/skipper:v0.21.206 skipper --help

Docker image

Docker image is available in Zalando's Open Source registry:

docker run -it registry.opensource.zalan.do/teapot/skipper:v0.21.206 skipper --help

v0.21.205

24 Sep 13:48
ce501f8
Compare
Choose a tag to compare

Changes

skipper: add server connection keepalive limits (#3246)

Clients may connect to a subset of Skipper fleet which leads to uneven
request distribution and increased cpu usage.

Autoscaling of Skipper fleet is not effective because clients stay
connected to old instances while new instances are underutilized.

This change adds ConnManager that tracks creation of new connections
and closes connections when their age or number of requests served
reaches configured limits.

Multiarch Docker image

Multiarch Docker image is available in Github's docker registry:

docker run -it ghcr.io/zalando/skipper:v0.21.205 skipper --help

Docker image

Docker image is available in Zalando's Open Source registry:

docker run -it registry.opensource.zalan.do/teapot/skipper:v0.21.205 skipper --help

v0.21.204

23 Sep 12:59
cbc392a
Compare
Choose a tag to compare

Changes

build(deps): bump oss-fuzz-base/base-builder-go in /.clusterfuzzlite (#3242)

Bumps oss-fuzz-base/base-builder-go from 130bed2 to 2347618.


updated-dependencies:

  • dependency-name: oss-fuzz-base/base-builder-go
    dependency-type: direct:production
    ...

Multiarch Docker image

Multiarch Docker image is available in Github's docker registry:

docker run -it ghcr.io/zalando/skipper:v0.21.204 skipper --help

Docker image

Docker image is available in Zalando's Open Source registry:

docker run -it registry.opensource.zalan.do/teapot/skipper:v0.21.204 skipper --help

v0.21.203

20 Sep 17:49
e34f988
Compare
Choose a tag to compare

Changes

filters/auth: release tokeninfo cache mutex earlier (#3241)

  • filters/auth: reset parallelism for BenchmarkTokeninfoCache

  • filters/auth: release tokeninfo cache mutex earlier

Cached value is readonly so release cache mutex early to
copy and adjust cached value outside of critical section.

goos: linux
goarch: amd64
pkg: github.com/zalando/skipper/filters/auth
cpu: Intel(R) Core(TM) i5-8350U CPU @ 1.70GHz
                                                  │    HEAD~1    │                 HEAD                 │
                                                  │    sec/op    │    sec/op     vs base                │
TokeninfoCache/tokens=1,cacheSize=1,p=0-8           639.4n ± 21%   271.9n ±  6%  -57.48% (p=0.000 n=10)
TokeninfoCache/tokens=2,cacheSize=2,p=0-8           625.7n ± 24%   420.4n ± 32%  -32.81% (p=0.000 n=10)
TokeninfoCache/tokens=100,cacheSize=100,p=0-8       869.8n ± 17%   463.4n ±  3%  -46.72% (p=0.000 n=10)
TokeninfoCache/tokens=100,cacheSize=100,p=10000-8   670.8n ± 12%   594.6n ±  2%  -11.36% (p=0.000 n=10)
TokeninfoCache/tokens=4,cacheSize=2,p=0-8           2.553m ±  1%   2.554m ±  1%        ~ (p=0.684 n=10)
TokeninfoCache/tokens=100,cacheSize=10,p=0-8        2.581m ±  1%   2.566m ±  0%        ~ (p=0.089 n=10)
geomean                                             10.74µ         7.688µ        -28.44%

                                                  │   HEAD~1   │                HEAD                 │
                                                  │    B/op    │    B/op     vs base                 │
TokeninfoCache/tokens=1,cacheSize=1,p=0-8           344.0 ± 0%   344.0 ± 0%       ~ (p=1.000 n=10) ¹
TokeninfoCache/tokens=2,cacheSize=2,p=0-8           344.0 ± 0%   344.0 ± 0%       ~ (p=1.000 n=10) ¹
TokeninfoCache/tokens=100,cacheSize=100,p=0-8       344.0 ± 0%   344.0 ± 0%       ~ (p=1.000 n=10) ¹
TokeninfoCache/tokens=100,cacheSize=100,p=10000-8   366.0 ± 0%   366.0 ± 1%       ~ (p=0.452 n=10)
TokeninfoCache/tokens=4,cacheSize=2,p=0-8           27.00 ± 4%   27.00 ± 0%       ~ (p=0.474 n=10)
TokeninfoCache/tokens=100,cacheSize=10,p=0-8        28.00 ± 7%   27.00 ± 7%       ~ (p=0.259 n=10)
geomean                                             149.7        148.8       -0.60%
¹ all samples are equal

                                                  │    HEAD~1    │                HEAD                 │
                                                  │  allocs/op   │ allocs/op   vs base                 │
TokeninfoCache/tokens=1,cacheSize=1,p=0-8           3.000 ± 0%     3.000 ± 0%       ~ (p=1.000 n=10) ¹
TokeninfoCache/tokens=2,cacheSize=2,p=0-8           3.000 ± 0%     3.000 ± 0%       ~ (p=1.000 n=10) ¹
TokeninfoCache/tokens=100,cacheSize=100,p=0-8       3.000 ± 0%     3.000 ± 0%       ~ (p=1.000 n=10) ¹
TokeninfoCache/tokens=100,cacheSize=100,p=10000-8   3.000 ± 0%     3.000 ± 0%       ~ (p=1.000 n=10) ¹
TokeninfoCache/tokens=4,cacheSize=2,p=0-8           0.000 ± 0%     0.000 ± 0%       ~ (p=1.000 n=10) ¹
TokeninfoCache/tokens=100,cacheSize=10,p=0-8        0.000 ± 0%     0.000 ± 0%       ~ (p=1.000 n=10) ¹
geomean                                                        ²               +0.00%                ²
¹ all samples are equal
² summaries must be >0 to compute geomean

Multiarch Docker image

Multiarch Docker image is available in Github's docker registry:

docker run -it ghcr.io/zalando/skipper:v0.21.203 skipper --help

Docker image

Docker image is available in Zalando's Open Source registry:

docker run -it registry.opensource.zalan.do/teapot/skipper:v0.21.203 skipper --help

v0.21.202

19 Sep 20:36
d161ac5
Compare
Choose a tag to compare

Changes

filters/auth: fix assert usage in tokeninfocache tests (#3240)

Second argument of assert.Equal should be expected value.

Multiarch Docker image

Multiarch Docker image is available in Github's docker registry:

docker run -it ghcr.io/zalando/skipper:v0.21.202 skipper --help

Docker image

Docker image is available in Zalando's Open Source registry:

docker run -it registry.opensource.zalan.do/teapot/skipper:v0.21.202 skipper --help

v0.21.201

19 Sep 18:00
7f6f037
Compare
Choose a tag to compare

Changes

predicates/host: add HostAny benchmark (#3239)

Current implementation uses linear probe that works well for small number
of hosts.

I thought about implementing it using map lookup but that is slower for
less than 50 hosts so I propose to only keep the benchmark that may be used
for future improvements.

Here is comparison with map-based implementation for reference:

              │    HEAD~1    │                 HEAD                  │
              │    sec/op    │    sec/op      vs base                │
HostAny/1-8      3.738n ± 3%    5.844n ±  3%  +56.36% (p=0.000 n=10)
HostAny/2-8      7.728n ± 4%    9.274n ±  2%  +20.02% (p=0.000 n=10)
HostAny/5-8      11.32n ± 4%    13.09n ±  4%  +15.63% (p=0.000 n=10)
HostAny/10-8     22.35n ± 3%    11.86n ± 15%  -46.94% (p=0.000 n=10)
HostAny/20-8     7.684n ± 2%   11.310n ±  9%  +47.19% (p=0.000 n=10)
HostAny/50-8     66.75n ± 4%    12.10n ± 12%  -81.87% (p=0.000 n=10)
HostAny/100-8   161.80n ± 1%    11.92n ± 13%  -92.63% (p=0.000 n=10)
geomean          17.98n         10.46n        -41.79%

              │    HEAD~1    │                HEAD                 │
              │     B/op     │    B/op     vs base                 │
HostAny/1-8     0.000 ± 0%     0.000 ± 0%       ~ (p=1.000 n=10) ¹
HostAny/2-8     0.000 ± 0%     0.000 ± 0%       ~ (p=1.000 n=10) ¹
HostAny/5-8     0.000 ± 0%     0.000 ± 0%       ~ (p=1.000 n=10) ¹
HostAny/10-8    0.000 ± 0%     0.000 ± 0%       ~ (p=1.000 n=10) ¹
HostAny/20-8    0.000 ± 0%     0.000 ± 0%       ~ (p=1.000 n=10) ¹
HostAny/50-8    0.000 ± 0%     0.000 ± 0%       ~ (p=1.000 n=10) ¹
HostAny/100-8   0.000 ± 0%     0.000 ± 0%       ~ (p=1.000 n=10) ¹
geomean                    ²               +0.00%                ²
¹ all samples are equal
² summaries must be >0 to compute geomean

              │    HEAD~1    │                HEAD                 │
              │  allocs/op   │ allocs/op   vs base                 │
HostAny/1-8     0.000 ± 0%     0.000 ± 0%       ~ (p=1.000 n=10) ¹
HostAny/2-8     0.000 ± 0%     0.000 ± 0%       ~ (p=1.000 n=10) ¹
HostAny/5-8     0.000 ± 0%     0.000 ± 0%       ~ (p=1.000 n=10) ¹
HostAny/10-8    0.000 ± 0%     0.000 ± 0%       ~ (p=1.000 n=10) ¹
HostAny/20-8    0.000 ± 0%     0.000 ± 0%       ~ (p=1.000 n=10) ¹
HostAny/50-8    0.000 ± 0%     0.000 ± 0%       ~ (p=1.000 n=10) ¹
HostAny/100-8   0.000 ± 0%     0.000 ± 0%       ~ (p=1.000 n=10) ¹
geomean                    ²               +0.00%                ²
¹ all samples are equal
² summaries must be >0 to compute geomean

Multiarch Docker image

Multiarch Docker image is available in Github's docker registry:

docker run -it ghcr.io/zalando/skipper:v0.21.201 skipper --help

Docker image

Docker image is available in Zalando's Open Source registry:

docker run -it registry.opensource.zalan.do/teapot/skipper:v0.21.201 skipper --help

v0.21.200

19 Sep 16:34
3c2a615
Compare
Choose a tag to compare

Changes

Add configurable cookie name prefix for the OIDC filter (#3234)

Allow specifying cookie name prefix for OIDC filters

Multiarch Docker image

Multiarch Docker image is available in Github's docker registry:

docker run -it ghcr.io/zalando/skipper:v0.21.200 skipper --help

Docker image

Docker image is available in Zalando's Open Source registry:

docker run -it registry.opensource.zalan.do/teapot/skipper:v0.21.200 skipper --help