Commit Graph
39 Commits
Author SHA1 Message Date
Nikola JokicandCopilot App 7833438dd7 Handle scale set messages with parallel workers in the listener
The scaleset listener no longer dissects the message it polls. It owns
session management, polling and acking, and hands the whole message to a
single Scale call, so acquiring jobs and recording metrics move to the
only component that still reads them.

That handover is what makes the work parallelisable. The listener used to
replay a message one API call at a time, in a fixed order: every job
started patch, then every job completed, then the scale patch. The job
events touch distinct EphemeralRunners and carry no ordering between
them, so they now run across a bounded worker pool, with the worker that
patches the EphemeralRunnerSet running alongside them.

The one ordering that does matter is kept. deleteIdleEphemeralRunners
skips a runner only once it carries a job request ID, so a patch that
lowers the replica count could offer up a runner that just picked up a
job if it were published while job started patches were still in flight.
The scaling worker therefore waits for the event workers on a scale down,
and only then. A patch that scales up or holds cannot delete anything.

Kubernetes has no bulk write: get, create, update, patch and delete are
single-resource verbs, and deletecollection is the only collection-scoped
mutating verb there is, so N events cannot be collapsed into fewer
requests. Server side apply is a PATCH on one object URI and does not
change that. Issuing the N patches concurrently over the one HTTP/2
connection is the available win; the alternative is writing fewer
objects, which trades N cheap independent writes for one contended,
watch-amplifying, size-bounded write.

The pool size is configurable through listenerConfig.scaler.workers,
alongside the existing qps and burst, and defaults to 10: two calls per
event keeps a full pool well inside the default QPS budget.

Statistics are now cached by the scaler. The listener stopped tracking
them, and a long poll that times out carries no message at all, so
without the cache an idle scale set would stop converging.

Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
2026-09-18 15:05:54 +02:00
Nikola Jokic 7c68e1d318 Set listener qps and burst (#4558) 2026-09-08 15:49:58 +02:00
Nikola Jokic 802dc28d38 Add multi-label support to scalesets (#4408) 2026-03-19 15:29:40 +01:00
Nikola Jokic 40595d806f Add chart-level API to customize internal resources (#4410) 2026-03-18 14:44:30 +01:00
Marcus Ramberg cac7a40b70 Add support for giving kubernetes mode scaleset service account additional permissions (#4282) 2025-11-21 15:56:08 +01:00
Nikola JokicandCopilot e3ed1ba226 Introduce new kubernetes-novolume mode (#4250)
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
2025-10-03 12:03:38 +02:00
Yusuke Kuoka f731873df9 Add workflow name and target labels (#4240) 2025-09-30 16:01:51 +02:00
Cory Calahan ad826725ce Update example GitHub URLs in values.yaml to include an example for enterprise account-level runners (#4181) 2025-07-16 10:40:38 +02:00
Mark Huijgen 0b2534ebc9 Fix dind sidecar template (#4128) 2025-06-16 12:14:18 +02:00
Jeev B e858d67926 Fix indentation of startupProbe attributes in dind sidecar (#4126) 2025-06-14 21:05:53 +02:00
Nikola Jokic e46c929241 Azure Key Vault integration to resolve secrets (#4090) 2025-06-11 15:53:33 +02:00
Nikola Jokic 9d8c59aeb3 Add startup probe to dind side-car (#4117) 2025-06-11 15:50:30 +02:00
Ryo Sakamoto 97697e80b4 Add job_workflow_ref label to listener metrics (#4054)
Signed-off-by: rskmm0chang <rskmm0chang@hatena.ne.jp>
2025-06-05 08:33:30 +02:00
Nikola Jokic 1dbb88cb9e Allow use of client id as an app id (#4057) 2025-05-16 16:21:06 +02:00
Nikola Jokic 5a960b5ebb Create configurable metrics (#3975) 2025-03-24 15:27:42 +01:00
J. FernándezandNikola Jokic 3c1a323381 feat: allow namespace overrides (#3797)
Signed-off-by: Jesús Fernández <7312236+fernandezcuesta@users.noreply.github.com>
Co-authored-by: Nikola Jokic <jokicnikola07@gmail.com>
2025-03-18 21:41:04 +01:00
Nikola Jokic 87938ee5bf Include custom annotations and labels to all resources created by gha-runner-scale-set chart (#3934) 2025-03-07 11:57:48 +01:00
Cees-Jan Kiewiet 2f5c981d46 Drop verbose flag from runner scale set init-dind-externals copy (#3805) 2025-03-05 21:02:27 +01:00
John Wesley Walker IIIandBassem Dghaidi 790191e987 Clarify syntax for githubConfigSecret (#3812)
Co-authored-by: Bassem Dghaidi <568794+Link-@users.noreply.github.com>
2025-01-21 13:29:41 +01:00
nasa9084 98854ef9c0 Fix doc comment for listenerTemplate (#3436) 2024-04-15 11:48:30 +02:00
Nikola Jokic 46cfbb6ec7 Fix documented dind expansion (#3368) 2024-03-19 15:24:58 +01:00
Nikola Jokic 2c0e53951b Fix tests and comment string for docker socket mounted path (#3366) 2024-03-19 11:29:07 +01:00
Nikola Jokic f7eb88ce9c Change minRunners behavior and fix the new listener min runners (#3139) 2023-12-13 19:39:21 +01:00
kahirokunn e1edb84abe chore: Service accounts in Kubernetes mode can now be annotated. (#2566)
Signed-off-by: kahirokunn <okinakahiro@gmail.com>
2023-10-18 13:37:39 +02:00
Dmitry ChepurovskiyandBassem Dghaidi 16666e1bba Fix #2809 : replace TLS dockerd connection with unix socket (#2833)
Co-authored-by: Bassem Dghaidi <568794+Link-@users.noreply.github.com>
2023-09-22 12:41:50 +02:00
Nikola JokicandBassem Dghaidi ea2fb32e20 Extend and generate crds allowing listener pod spec change (#2758)
Co-authored-by: Bassem Dghaidi <568794+Link-@users.noreply.github.com>
2023-09-14 15:33:29 +02:00
Andi Büchler 564c112b1a Fix trivial typos (#2856) 2023-09-07 13:51:41 +02:00
Francesco Virga c7dce2bbb7 Documenting the runner container command in values.yaml (#2854) 2023-09-07 13:47:08 +02:00
Nikola Jokic 1c360d7e26 Document customization for containerModes (#2777) 2023-08-18 11:03:28 +02:00
Tingluo Huang e45e4c53f1 Add E2E test to assert self-signed CA support. (#2458) 2023-03-31 10:31:25 -04:00
Francesco Renzi c8e3bb5ec3 Remove containerMode from values (#2442) 2023-03-28 10:16:38 +01:00
Tingluo Huang 064039afc0 Ignore extra dind container when contaerinMode.type=dind. (#2418) 2023-03-17 09:26:51 +01:00
Bassem Dghaidi c465ace8fb Update the values.yaml sample for improved clarity (#2416) 2023-03-16 11:02:18 -04:00
Tingluo Huang 40811ebe0e Support the controller to watching a single namespace. (#2374) 2023-03-14 10:52:25 -04:00
Tingluo Huang 9e6c7d019f Delay role/rolebinding creation to gha-runner-scale-set installation time (#2363) 2023-03-14 09:45:44 -04:00
Tingluo Huang d7b589bed5 Helm chart react changes for the new runner image. (#2348) 2023-03-10 11:18:21 +00:00
c569304271 Add support for self-signed CA certificates (#2268)
Co-authored-by: Bassem Dghaidi <568794+Link-@users.noreply.github.com>
Co-authored-by: Nikola Jokic <jokicnikola07@gmail.com>
Co-authored-by: Tingluo Huang <tingluohuang@github.com>
2023-03-09 17:23:32 +00:00
Chris PattersonandTingluoHuang 41f2ca3ed9 Adding parameter to configure the runner set name. (#2279)
Co-authored-by: TingluoHuang <TingluoHuang@github.com>
2023-03-03 08:36:14 -05:00
Bassem DghaidiandAva S adad3d5530 Rename actions-runner-controller-2 and auto-scaling-runner-set helm charts (#2333)
Co-authored-by: Ava S <avastancu@github.com>
2023-03-01 07:16:03 -05:00