[{"data":1,"prerenderedAt":790},["ShallowReactive",2],{"/en-us/blog/how-we-removed-all-502-errors-by-caring-about-pid-1-in-kubernetes":3,"navigation-en-us":35,"banner-en-us":435,"footer-en-us":445,"blog-post-authors-en-us-Steve Azzopardi":687,"blog-related-posts-en-us-how-we-removed-all-502-errors-by-caring-about-pid-1-in-kubernetes":701,"assessment-promotions-en-us":741,"next-steps-en-us":780},{"id":4,"title":5,"authorSlugs":6,"body":8,"categorySlug":9,"config":10,"content":14,"description":8,"extension":24,"isFeatured":12,"meta":25,"navigation":26,"path":27,"publishedDate":20,"seo":28,"stem":32,"tagSlugs":33,"__hash__":34},"blogPosts/en-us/blog/how-we-removed-all-502-errors-by-caring-about-pid-1-in-kubernetes.yml","How We Removed All 502 Errors By Caring About Pid 1 In Kubernetes",[7],"steve-azzopardi",null,"engineering",{"slug":11,"featured":12,"template":13},"how-we-removed-all-502-errors-by-caring-about-pid-1-in-kubernetes",false,"BlogPost",{"title":15,"description":16,"authors":17,"heroImage":19,"date":20,"body":21,"category":9,"tags":22},"How we reduced 502 errors by caring about PID 1 in Kubernetes","For every deploy, scale down event, or pod termination, users of GitLab's Pages service were experiencing 502 errors. This explains how we found the root cause and rolled out a fix for it.",[18],"Steve Azzopardi","https://res.cloudinary.com/about-gitlab-com/image/upload/v1749682305/Blog/Hero%20Images/KubeCon2022.jpg","2022-05-17","\n\n_This blog post and linked pages contain information related to upcoming products, features, and functionality. It is important to note that the information presented is for informational purposes only. Please do not rely on this information for purchasing or planning purposes. As with all projects, the items mentioned in this blog post and linked pages are subject to change or delay. The development, release, and timing of any products, features, or functionality remain at the sole discretion of GitLab Inc._\n\nOur [SRE on call](https://handbook.gitlab.com/handbook/engineering/infrastructure/incident-management/#engineer-on-call-eoc-responsibilities)\nwas getting paged daily that one of our\n[SLIs](https://www.youtube.com/watch?v=tEylFyxbDLE) was\nburning through our\n[SLOs](https://www.youtube.com/watch?v=tEylFyxbDLE) for the [GitLab\nPages](https://docs.gitlab.com/ee/user/project/pages/) service. It was\nintermittent and short-lived, but enough to cause user-facing impact which we\nweren't comfortable with. This turned into alert fatigue because there wasn't\nenough time for the SRE on call to investigate the issue and it wasn't\nactionable since it recovered on its own.\n\nWe decided to open up an [investigation issue](https://gitlab.com/gitlab-com/gl-infra/reliability/-/issues/15497)\nfor these alerts. We had to find out what the issue was since we were\nshowing `502` errors to our users and we needed a\n[DRI](https://handbook.gitlab.com/handbook/people-group/directly-responsible-individuals/)\nthat wasn't on call to investigate.\n\n## What is even going on?\n\nAs an [SRE](https://handbook.gitlab.com/job-families/engineering/infrastructure/site-reliability-engineer/)\nat GitLab, you get to touch a lot of services that you didn't build yourself and\ninteract with system dependencies that you might have not touched before.\nThere's always detective work to do!\n\nWhen we looked at the GitLab Pages logs we found that it's always returning\n[`ErrDomainDoesNotExist`](https://gitlab.com/gitlab-org/gitlab-pages/-/blob/e1f1effa23c520d3b8b717d831ccab7ba3dd494f/internal/routing/middleware.go#L22-26)\nerrors which result in a `502` error to our users. GitLab Pages [sends a request](https://gitlab.com/gitlab-org/gitlab-pages/-/blob/e1f1effa23c520d3b8b717d831ccab7ba3dd494f/internal/source/gitlab/client/client.go#L101-127)\nto [GitLab Workhorse](https://docs.gitlab.com/ee/development/workhorse/),\nspecifically the `/api/v4/internal/pages` route.\n\nGitLab Workhorse is a Go service in front of our Ruby on Rails monolith and\nit's deployed as a [sidecar](https://www.magalix.com/blog/the-sidecar-pattern)\ninside of the `webservice pod`, which runs Ruby on Rails using the `Puma` web\nserver.\n\nWe used the internal IP to correlate the GitLab Pages requests with GitLab Workhorse\ncontainers. We looked at multiple requests and found that all the 502 requests\nhad the following error attached to them: [`502 Bad Gateway with dial tcp 127.0.0.1:8080: connect: connection refused`](https://gitlab.com/gitlab-org/gitlab/-/blob/f64be48cc737f5d12c1c30f724af540a836dcc94/workhorse/internal/badgateway/roundtripper.go#L43).\nThis means that GitLab Workhorse couldn't connect to the Puma web server. So we\nneeded to go another layer deeper.\n\nThe Puma web server is what runs the Ruby on Rails monolith which has an\ninternal API endpoint but Puma was never getting these requests since it wasn't\nrunning. What this tells us is that Kubernetes kept our pod in the\n[service](https://kubernetes.io/docs/concepts/services-networking/service/)\neven when Puma wasn't responding, despite having [readiness probes](https://gitlab.com/gitlab-org/charts/gitlab/-/blob/4bb638bccc6a676f9fdd5bbf800f7d2b977efd55/charts/gitlab/charts/webservice/templates/deployment.yaml#L279-287)\nconfigured.\n\nBelow is the request flow between GitLab Pages, GitLab Workhorse, and Puma/Webservice to try and make it more clear:\n\n![overview of the request flow](https://about.gitlab.com/images/blogimages/how-we-removed-all-502-errors-by-caring-about-pid-1-in-kubernetes/overview.png){: .shadow.center}\n\n## Attempt 1: Red herring\n\nWe shifted our focus on GitLab Workhorse and Puma to try and understand how\nGitLab Workhorse was returning 502 errors in the first place. We found some\n`502 Bad Gateway with dial tcp 127.0.0.1:8080: connect: connection refused`\nerrors during container startup time. How could this be? With the readiness\nprobe, the pod shouldn't be added to the\n[Endpoint](https://kubernetes.io/docs/concepts/services-networking/service/#over-capacity-endpoints)\nuntil [all readiness probes pass](https://gitlab.com/gitlab-com/gl-infra/reliability/-/issues/15497#note_899321775).\nWe later found out that it's because of a [polling\nmechanisim](https://gitlab.com/gitlab-com/gl-infra/reliability/-/issues/15497#note_899629314)\nthat we have for [Geo](https://docs.gitlab.com/ee/administration/geo/) which\nruns in the background, using a Goroutine in GitLab Workhorse, and pings Puma for Geo information.\nWe don't have Geo enabled on GitLab.com so we [simply disabled it](https://gitlab.com/gitlab-com/gl-infra/k8s-workloads/gitlab-com/-/merge_requests/1670)\nto reduce the noise.\n\nWe removed the 502 errors, but not the ones we want, just a red herring.\n\n## Attempt 2: Close but not quite\n\nAt this time, we were still burning through our SLO from time to time, so this\nwas still an urgent thing that we needed to fix. Now that we had cleaner logs for\n`502` errors it started to become a bit clearer that this is happening on pod\ntermination:\n\n```text\n2022-04-05 06:03:49.000 UTC: Readiness probe failed\n2022-04-05 06:03:51.000 UTC: Puma (127.0.0.1:8080) started shutdown.\n2022-04-05 06:04:04.526 UTC: Puma shutdown finished.\n2022-04-05 06:04:04.000 UTC - 2022-04-05 06:04:46.000 UTC: workhorse started serving 502 constantly.  42 seconds of serving 502 requests for any request that comes in apart from /api/v4/jobs/request\n```\n\nIn the timeline shown above, we see that we've kept serving requests well after\nour `Puma`/`webservice` container exited, and the first readiness probe failed.\nIf we look at the readiness probes we had on that pod we see the following:\n\n```shell\n$ kubectl -n gitlab get po gitlab-webservice-api-785cb54bbd-xpln2 -o jsonpath='{range .spec.containers[*]} {@.name \":\\n\\tliveness:\"} {@.livenessProbe} {\"\\n\\treadiness:\"} {@.readinessProbe} {\"\\n\"} {end}'\n webservice:\n        liveness: {\"failureThreshold\":3,\"httpGet\":{\"path\":\"/-/liveness\",\"port\":8080,\"scheme\":\"HTTP\"},\"initialDelaySeconds\":20,\"periodSeconds\":60,\"successThreshold\":1,\"timeoutSeconds\":30}\n        readiness: {\"failureThreshold\":3,\"httpGet\":{\"path\":\"/-/readiness\",\"port\":8080,\"scheme\":\"HTTP\"},\"initialDelaySeconds\":60,\"periodSeconds\":10,\"successThreshold\":1,\"timeoutSeconds\":2}\n  gitlab-workhorse:\n        liveness: {\"exec\":{\"command\":[\"/scripts/healthcheck\"]},\"failureThreshold\":3,\"initialDelaySeconds\":20,\"periodSeconds\":60,\"successThreshold\":1,\"timeoutSeconds\":30}\n        readiness: {\"exec\":{\"command\":[\"/scripts/healthcheck\"]},\"failureThreshold\":3,\"periodSeconds\":10,\"successThreshold\":1,\"timeoutSeconds\":2}\n\n```\n\nThis meant that for the `webservice` pod to be marked unhealthy and removed\nfrom the endpoints, Kubernetes had to get 3 consecutive failures with an\ninterval of 10 seconds, so in total that's 30 seconds. That seems a bit slow.\n\nOur next logical step was to reduce the `periodSeconds` for the readiness probe\nfor the `webservice` pod so we don't wait 30 seconds before removing the pod\nfrom the service when it becomes unhealthy.\n\nBefore doing so we had to understand if sending more requests to `/-/readiness`\nendpoint would have any knock-on effect with using more memory or anything\nelse. We had to [understand what the `/-/readiness` endpoint was doing](https://gitlab.com/gitlab-com/gl-infra/reliability/-/issues/15497#note_903812722)\nand if it was safe to increase the frequency at which we send requests. We\ndecided it was safe, and after enabling it on\n[staging](https://gitlab.com/gitlab-com/gl-infra/k8s-workloads/gitlab-com/-/merge_requests/1686#note_903877755),\nand\n[canary](https://gitlab.com/gitlab-com/gl-infra/k8s-workloads/gitlab-com/-/merge_requests/1688#note_904501848)\nwe didn't see any increase in CPU/Memory usage, as expected, and saw an\nimprovement in the removal of 502 errors, which made us more confident that\nthis was the issue. We rolled this out to\n[production](https://gitlab.com/gitlab-com/gl-infra/k8s-workloads/gitlab-com/-/merge_requests/1689)\nwith high hopes.\n\nAs usual, Production is a different story than Staging or Canary, and it showed\nthat it didn't remove all the 502 errors, just [enough to stop triggering the SLO every day](https://gitlab.com/gitlab-com/gl-infra/reliability/-/issues/15497#note_905993144),\nbut at least we removed the alert fatigue on the SRE on call. We were close, but not quite.\n\n## Attempt 3: All gone!\n\nAt this point, we were a bit lost and weren't sure what to look at next. We had\na bit of tunnel vision and kept focusing/blaming that we aren't removing the\nPod from the `Endpoint` quickly enough. We even looked at [Google Cloud Platform\nNEGs](https://cloud.google.com/kubernetes-engine/docs/how-to/standalone-neg) to\nsee if we could have faster readiness probes and remove the pod quicker. However,\nthis wasn't ideal [because we wouldn't have solved this for our self-hosting customers](https://gitlab.com/gitlab-com/gl-infra/reliability/-/issues/15497#note_908359286)\nwhich seem to be facing the same [problem](https://gitlab.com/gitlab-org/charts/gitlab/-/issues/2943).\n\nWhile researching we also came across a known problem with [running `Puma` in\nKubernetes](https://github.com/puma/puma/blob/bf2548ce300c2b4f671582bc756dcec5861e815f/docs/kubernetes.md),\nand thought that might be the solution. However, we already implemented a\n[blackout window](https://gitlab.com/gitlab-org/charts/gitlab/-/blob/c1b63f3a4867886bc1212d86985fc70e66b717c5/charts/gitlab/charts/webservice/templates/deployment.yaml#L223-224)\njust for this specific reason, so it couldn't be that either...in other words, it was another dead end.\n\nWe took a step back and looked at the [timelines one more time](https://gitlab.com/gitlab-com/gl-infra/reliability/-/issues/15497#note_910106152)\nand then it hit us. The Puma/webservice container is terminating within a\nfew seconds, but the GitLab Workhorse one is always taking 30 seconds. Is it because\nof the [long polling from GitLab Runner](https://gitlab.com/gitlab-org/gitlab-foss/-/issues/21698)? 30 seconds\nis a \"special\" number for Kubernetes [pod termination](https://kubernetes.io/docs/concepts/workloads/pods/pod-lifecycle/#pod-termination).\nWhen Kubernetes deletes a pod it firsts sends the `TERM` signal to the\ncontainer and waits 30 seconds, if the container hasn't exited yet, it will\nsend a `KILL` signal. This indicated that maybe GitLab Workhorse was never\nshutting down and Kubernetes had to kill it.\n\nOnce more we looked at GitLab Workhorse source code and [searched for the `SIGTERM` usage](https://gitlab.com/gitlab-org/gitlab/-/blob/d66f10e169a08cedcbfe70e3ea46cbfbb20d972d/workhorse/main.go#L238-258)\nand it did seem to support [graceful termination](https://gitlab.com/gitlab-org/gitlab/-/merge_requests/62701) and\nit also had explicit logic about long polling requests, so is this just another\ndead end? Luckily when the `TERM` signal is sent, Workhorse [logs a message that\nit's shutting down](https://gitlab.com/gitlab-org/gitlab/-/merge_requests/62701). We looked\nat our logs for this specific message and didn't see anything. Is this it? We\naren't gracefully shutting down? But how? Why does it result in 502 errors?\nWhy do the GitLab Pages keep using the same pod that is terminating?\n\nWe know that the `TERM` signal is being sent to PID 1 inside of the container,\nand that process should handle the `TERM` signal for graceful shutdown. We\nlooked at the GitLab Workhorse process tree and this is what we found:\n\n```sh\ngit@gitlab-webservice-default-5d85b6854c-sbx2z:/$ ps faux\nUSER         PID %CPU %MEM    VSZ   RSS TTY      STAT START   TIME COMMAND\nroot        1015  0.0  0.0 805036  4588 ?        Rsl  13:12   0:00 runc init\ngit         1005  0.3  0.0   5992  3784 pts/0    Ss   13:12   0:00 bash\ngit         1014  0.0  0.0   8592  3364 pts/0    R+   13:12   0:00  \\_ ps faux\ngit            1  0.0  0.0   2420   532 ?        Ss   12:52   0:00 /bin/sh -c /scripts/start-workhorse\ngit           16  0.0  0.0   5728  3408 ?        S    12:52   0:00 /bin/bash /scripts/start-workhorse\ngit           19  0.0  0.3 1328480 33080 ?       Sl   12:52   0:00  \\_ gitlab-workhorse -logFile stdout -logFormat json -listenAddr 0.0.0.0:8181 -documentRoot /srv/gitlab/public -secretPath /etc/gitlab/gitlab-workhorse/secret -config /srv/gitlab/config/workhorse-config.toml\n```\n\nBingo! `gitlab-workhorse` is PID 19 in this case, and a child process of a\n[script](https://gitlab.com/gitlab-org/build/CNG/-/blob/92d3e22e9ff6c5cbb685aeea99813751d5e19a9d/gitlab-workhorse/Dockerfile#L51)\nthat we invoke. Taking a close look at the\n[script](https://gitlab.com/gitlab-org/build/CNG/-/blob/92d3e22e9ff6c5cbb685aeea99813751d5e19a9d/gitlab-workhorse/scripts/start-workhors)\nwe check if it listens to `TERM` and it doesn't! So far everything indicated\nthat GitLab Workhorse was never getting the `TERM` signal which ended up in receiving\n`KILL` after 30 seconds. We updated our `scripts/start-workhorse` to use\n[`exec(1)`](https://linux.die.net/man/1/exec) so that `gitlab-workhorse`\nreplaced the PID of our bash script, that should have worked, right? When we tested\nthis locally we then saw the following process tree.\n\n```shell\ngit@gitlab-webservice-default-84c68fc9c9-xcsnm:/$ ps faux\nUSER         PID %CPU %MEM    VSZ   RSS TTY      STAT START   TIME COMMAND\ngit          167  0.0  0.0   5992  3856 pts/0    Ss   14:27   0:00 bash\ngit          181  0.0  0.0   8592  3220 pts/0    R+   14:27   0:00  \\_ ps faux\ngit            1  0.0  0.0   2420   520 ?        Ss   14:24   0:00 /bin/sh -c /scripts/start-workhorse\ngit           17  0.0  0.3 1328228 32800 ?       Sl   14:24   0:00 gitlab-workhorse -logFile stdout -logFormat json -listenAddr 0.0.0.0:8181 -documentRoot /srv/gitlab/public -secretPath /etc/gitlab/gitlab-workhorse/secret -config /srv/gitlab/config/workhorse-config.toml\n```\n\nThis changed a bit: this shows that `gitlab-workhorse` was no longer a child\nprocess of `/scripts/start-workhorse` however `/bin/sh` was still PID 1. What is even\ninvoking `/bin/sh` that we didn't see anywhere in our\n[Dockerfile](https://gitlab.com/gitlab-org/build/CNG/-/blob/92d3e22e9ff6c5cbb685aeea99813751d5e19a9d/gitlab-workhorse/Dockerfile)?\nAfter some thumb-twiddling, we had an idea that the container runtime is invoking\n`/bin/sh`. We went back to basics and looked at the\n[`CMD`](https://docs.docker.com/engine/reference/builder/#cmd) documentation to\nsee if we were missing something, and we were. We read the following:\n\n> If you use the shell form of the CMD, then the \u003Ccommand> will execute in `/bin/sh -c`:\n>\n> ```\n> FROM ubuntu\n> CMD echo \"This is a test.\" | wc -\n> ```\n>\n> If you want to run your \u003Ccommand> without a shell then you must express the command as a JSON array and give the full path to the executable. This array form is the preferred format of CMD. Any additional parameters must be individually expressed as strings in the array:\n>\n> ```\n> FROM ubuntu\n> CMD [\"/usr/bin/wc\",\"--help\"]\n> ```\n\nThis was exactly [what we were doing](https://gitlab.com/gitlab-org/build/CNG/-/blob/92d3e22e9ff6c5cbb685aeea99813751d5e19a9d/gitlab-workhorse/Dockerfile#L51)!\nwe weren't using `CMD` in `exec form`, but in `shell form`. Changing this confirmed\nthat `gitlab-workhorse` is now PID 1, and also receives the termination signal\nafter testing it locally:\n\n```shell\ngit@gitlab-webservice-default-84c68fc9c9-lzwmp:/$ ps faux\nUSER         PID %CPU %MEM    VSZ   RSS TTY      STAT START   TIME COMMAND\ngit           65  1.0  0.0   5992  3704 pts/0    Ss   15:25   0:00 bash\ngit           73  0.0  0.0   8592  3256 pts/0    R+   15:25   0:00  \\_ ps faux\ngit            1  0.2  0.3 1328228 32288 ?       Ssl  15:24   0:00 gitlab-workhorse -logFile stdout -logFormat json -listenAddr 0.0.0.0:8181 -documentRoot /srv/gitlab/public -secretPath /etc/gitlab/gitlab-workhorse/secret -config /srv/gitlab/config/workhorse-config.toml\n```\n\n```text\n{\"level\":\"info\",\"msg\":\"shutdown initiated\",\"shutdown_timeout_s\":61,\"signal\":\"terminated\",\"time\":\"2022-04-13T15:27:57Z\"}\n{\"level\":\"info\",\"msg\":\"keywatcher: shutting down\",\"time\":\"2022-04-13T15:27:57Z\"}\n{\"error\":null,\"level\":\"fatal\",\"msg\":\"shutting down\",\"time\":\"2022-04-13T15:27:57Z\"}\n```\n\nOk, then we just needed to update `exec` and `CMD []` and we would have been\ndone, right? Almost. GitLab Workhorse proxies all of the requests for the API, Web, and Git requests so we couldn't just do a big change and expect that everything is going to be OK. We had to progressively roll this out to make\nsure we didn't break any existing working behavior since this affects all the\nrequests we get to GitLab.com. To do this, we hid it behind a [feature\nflag](https://gitlab.com/gitlab-org/build/CNG/-/merge_requests/972) so GitLab\nWorkhorse is only PID 1 when the `GITLAB_WORKHORSE_EXEC` environment variable\nis set. This allowed us to deploy the change and only enable it on a small part\nof our fleet to see if we see any problems. We were a bit more careful here and\nrolled it out zone by zone in Production since we run on 3 zones. When we\nrolled it out in the [first\nzone](https://gitlab.com/gitlab-com/gl-infra/reliability/-/issues/15497#note_919259030)\nwe saw all 502 errors disappear! After fully rolling this out we see that [the\nproblem is fixed and it had no negative side\neffects](https://gitlab.com/gitlab-com/gl-infra/reliability/-/issues/15497#note_920585707). Hurray!\n\nWe still had one question unanswered, why were GitLab Pages still trying to use\nthe same connection even after the Pod was removed from the Service because it was\nscheduled for deletion? When we looked at Go internals we see that [Go reuses\nTCP connections](https://gitlab.com/gitlab-com/gl-infra/reliability/-/issues/15497#note_920642770)\nif we close the body of the request. So even though it's not part of the Service\nwe can still keep the TCP connection open and send requests – this explains why\nwe kept seeing 502 on pod being terminated and always from the same GitLab\nPages pod.\n\nNow it's all gone!\n\n## More things that we can explore\n\n1. We've made graceful termination for GitLab Workhorse as [default](https://gitlab.com/gitlab-com/gl-infra/k8s-workloads/gitlab-com/-/merge_requests/1732).\n1. Audit all of our Dockerfiles that use `CMD command` and fix them. We've found 10, and [fixed all of them](https://gitlab.com/gitlab-org/charts/gitlab/-/issues/3249).\n1. [Better readiness Probe defaults for `webservice` pod](https://gitlab.com/gitlab-org/charts/gitlab/-/merge_requests/2518).\n1. Add [linting](https://gitlab.com/gitlab-org/charts/gitlab/-/issues/3253) for Dockerfiles.\n1. See if any of our child processes need [zombie process reaping](https://blog.phusion.nl/2015/01/20/docker-and-the-pid-1-zombie-reaping-problem/).\n\n## Takeaways\n\n1. We should always care about what is PID 1 in a container.\n1. Always try and use `CMD [\"executable\",\"param1\",\"param2\"]` in your Dockerfile.\n1. Pods are removed from the Service/Endpoint in async.\n1. If you are on GKE [NEGs](https://cloud.google.com/kubernetes-engine/docs/how-to/standalone-neg) might be better for readinessProbes.\n1. By default, there is a 30 second grace period between the `TERM` signal and the `KILL` signal when Pods terminate. You can update the time between the signals `terminationGracePeriodSeconds`.\n1. The Go `http.Client` will reuse the TCP connection if [the body is closed](https://cs.opensource.google/go/go/+/refs/tags/go1.18.2:src/net/http/response.go;l=59-64) which in this case made the issue worse.\n\nThank you to [@igorwwwwwwwwwwwwwwwwwwww](https://gitlab.com/igorwwwwwwwwwwwwwwwwwwww), [@gsgl](https://gitlab.com/gsgl), [@jarv](https://gitlab.com/jarv), and [@cmcfarland](https://gitlab.com/cmcfarland) for helping me debug this problem!\n\n",[23],"kubernetes","yml",{},true,"/en-us/blog/how-we-removed-all-502-errors-by-caring-about-pid-1-in-kubernetes",{"title":15,"description":16,"ogTitle":15,"ogDescription":16,"noIndex":12,"ogImage":19,"ogUrl":29,"ogSiteName":30,"ogType":31,"canonicalUrls":29},"https://about.gitlab.com/blog/how-we-removed-all-502-errors-by-caring-about-pid-1-in-kubernetes","https://about.gitlab.com","article","en-us/blog/how-we-removed-all-502-errors-by-caring-about-pid-1-in-kubernetes",[23],"uw7cskmKkZXwMGbX14uCisH1W_jedengjnMsahqmzW4",{"data":36},{"logo":37,"freeTrial":42,"sales":47,"login":52,"items":57,"search":365,"minimal":396,"duo":415,"pricingDeployment":425},{"config":38},{"href":39,"dataGaName":40,"dataGaLocation":41},"/","gitlab logo","header",{"text":43,"config":44},"Get free trial",{"href":45,"dataGaName":46,"dataGaLocation":41},"https://gitlab.com/-/trial_registrations/new?glm_source=about.gitlab.com&glm_content=default-saas-trial/","free trial",{"text":48,"config":49},"Talk to sales",{"href":50,"dataGaName":51,"dataGaLocation":41},"/sales/","sales",{"text":53,"config":54},"Sign in",{"href":55,"dataGaName":56,"dataGaLocation":41},"https://gitlab.com/users/sign_in/","sign in",[58,85,180,185,286,346],{"text":59,"config":60,"cards":62},"Platform",{"dataNavLevelOne":61},"platform",[63,69,77],{"title":59,"description":64,"link":65},"The intelligent orchestration platform for DevSecOps",{"text":66,"config":67},"Explore our Platform",{"href":68,"dataGaName":61,"dataGaLocation":41},"/platform/",{"title":70,"description":71,"link":72},"GitLab Duo Agent Platform","Agentic AI for the entire software lifecycle",{"text":73,"config":74},"Meet GitLab Duo",{"href":75,"dataGaName":76,"dataGaLocation":41},"/gitlab-duo-agent-platform/","gitlab duo agent platform",{"title":78,"description":79,"link":80},"Why GitLab","See the top reasons enterprises choose GitLab",{"text":81,"config":82},"Learn more",{"href":83,"dataGaName":84,"dataGaLocation":41},"/why-gitlab/","why gitlab",{"text":86,"left":26,"config":87,"link":89,"lists":93,"footer":162},"Product",{"dataNavLevelOne":88},"solutions",{"text":90,"config":91},"View all Solutions",{"href":92,"dataGaName":88,"dataGaLocation":41},"/solutions/",[94,118,141],{"title":95,"description":96,"link":97,"items":102},"Automation","CI/CD and automation to accelerate deployment",{"config":98},{"icon":99,"href":100,"dataGaName":101,"dataGaLocation":41},"AutomatedCodeAlt","/solutions/delivery-automation/","automated software delivery",[103,107,110,114],{"text":104,"config":105},"CI/CD",{"href":106,"dataGaLocation":41,"dataGaName":104},"/solutions/continuous-integration/",{"text":70,"config":108},{"href":75,"dataGaLocation":41,"dataGaName":109},"gitlab duo agent platform - product menu",{"text":111,"config":112},"Source Code Management",{"href":113,"dataGaLocation":41,"dataGaName":111},"/solutions/source-code-management/",{"text":115,"config":116},"Automated Software Delivery",{"href":100,"dataGaLocation":41,"dataGaName":117},"Automated software delivery",{"title":119,"description":120,"link":121,"items":126},"Security","Deliver code faster without compromising security",{"config":122},{"href":123,"dataGaName":124,"dataGaLocation":41,"icon":125},"/solutions/application-security-testing/","security and compliance","ShieldCheckLight",[127,131,136],{"text":128,"config":129},"Application Security Testing",{"href":123,"dataGaName":130,"dataGaLocation":41},"Application security testing",{"text":132,"config":133},"Software Supply Chain Security",{"href":134,"dataGaLocation":41,"dataGaName":135},"/solutions/supply-chain/","Software supply chain security",{"text":137,"config":138},"Software Compliance",{"href":139,"dataGaName":140,"dataGaLocation":41},"/solutions/software-compliance/","software compliance",{"title":142,"link":143,"items":148},"Measurement",{"config":144},{"icon":145,"href":146,"dataGaName":147,"dataGaLocation":41},"DigitalTransformation","/solutions/visibility-measurement/","visibility and measurement",[149,153,157],{"text":150,"config":151},"Visibility & Measurement",{"href":146,"dataGaLocation":41,"dataGaName":152},"Visibility and Measurement",{"text":154,"config":155},"Value Stream Management",{"href":156,"dataGaLocation":41,"dataGaName":154},"/solutions/value-stream-management/",{"text":158,"config":159},"Analytics & Insights",{"href":160,"dataGaLocation":41,"dataGaName":161},"/solutions/analytics-and-insights/","Analytics and insights",{"title":163,"items":164},"GitLab for",[165,170,175],{"text":166,"config":167},"Enterprise",{"href":168,"dataGaLocation":41,"dataGaName":169},"/enterprise/","enterprise",{"text":171,"config":172},"Small Business",{"href":173,"dataGaLocation":41,"dataGaName":174},"/small-business/","small business",{"text":176,"config":177},"Public Sector",{"href":178,"dataGaLocation":41,"dataGaName":179},"/solutions/public-sector/","public sector",{"text":181,"config":182},"Pricing",{"href":183,"dataGaName":184,"dataGaLocation":41,"dataNavLevelOne":184},"/pricing/","pricing",{"text":186,"config":187,"link":189,"lists":193,"feature":273},"Resources",{"dataNavLevelOne":188},"resources",{"text":190,"config":191},"View all resources",{"href":192,"dataGaName":188,"dataGaLocation":41},"/resources/",[194,227,245],{"title":195,"items":196},"Getting started",[197,202,207,212,217,222],{"text":198,"config":199},"Install",{"href":200,"dataGaName":201,"dataGaLocation":41},"/install/","install",{"text":203,"config":204},"Quick start guides",{"href":205,"dataGaName":206,"dataGaLocation":41},"/get-started/","quick setup checklists",{"text":208,"config":209},"Learn",{"href":210,"dataGaLocation":41,"dataGaName":211},"https://university.gitlab.com/","learn",{"text":213,"config":214},"Product documentation",{"href":215,"dataGaName":216,"dataGaLocation":41},"https://docs.gitlab.com/","product documentation",{"text":218,"config":219},"Best practice videos",{"href":220,"dataGaName":221,"dataGaLocation":41},"/getting-started-videos/","best practice videos",{"text":223,"config":224},"Integrations",{"href":225,"dataGaName":226,"dataGaLocation":41},"/integrations/","integrations",{"title":228,"items":229},"Discover",[230,235,240],{"text":231,"config":232},"Customer success stories",{"href":233,"dataGaName":234,"dataGaLocation":41},"/customers/","customer success stories",{"text":236,"config":237},"Blog",{"href":238,"dataGaName":239,"dataGaLocation":41},"/blog/","blog",{"text":241,"config":242},"Remote",{"href":243,"dataGaName":244,"dataGaLocation":41},"https://handbook.gitlab.com/handbook/company/culture/all-remote/","remote",{"title":246,"items":247},"Connect",[248,253,258,263,268],{"text":249,"config":250},"GitLab Services",{"href":251,"dataGaName":252,"dataGaLocation":41},"/services/","services",{"text":254,"config":255},"Community",{"href":256,"dataGaName":257,"dataGaLocation":41},"/community/","community",{"text":259,"config":260},"Forum",{"href":261,"dataGaName":262,"dataGaLocation":41},"https://forum.gitlab.com/","forum",{"text":264,"config":265},"Events",{"href":266,"dataGaName":267,"dataGaLocation":41},"/events/","events",{"text":269,"config":270},"Partners",{"href":271,"dataGaName":272,"dataGaLocation":41},"/partners/","partners",{"backgroundColor":274,"textColor":275,"text":276,"image":277,"link":281},"#2f2a6b","#fff","Insights for the future of software development",{"altText":278,"config":279},"the source promo card",{"src":280},"https://res.cloudinary.com/about-gitlab-com/image/upload/v1758208064/dzl0dbift9xdizyelkk4.svg",{"text":282,"config":283},"Read the latest",{"href":284,"dataGaName":285,"dataGaLocation":41},"/the-source/","the source",{"text":287,"config":288,"lists":290},"Company",{"dataNavLevelOne":289},"company",[291],{"items":292},[293,298,304,306,311,316,321,326,331,336,341],{"text":294,"config":295},"About",{"href":296,"dataGaName":297,"dataGaLocation":41},"/company/","about",{"text":299,"config":300,"footerGa":303},"Jobs",{"href":301,"dataGaName":302,"dataGaLocation":41},"/jobs/","jobs",{"dataGaName":302},{"text":264,"config":305},{"href":266,"dataGaName":267,"dataGaLocation":41},{"text":307,"config":308},"Leadership",{"href":309,"dataGaName":310,"dataGaLocation":41},"/company/team/e-group/","leadership",{"text":312,"config":313},"Team",{"href":314,"dataGaName":315,"dataGaLocation":41},"/company/team/","team",{"text":317,"config":318},"Handbook",{"href":319,"dataGaName":320,"dataGaLocation":41},"https://handbook.gitlab.com/","handbook",{"text":322,"config":323},"Investor relations",{"href":324,"dataGaName":325,"dataGaLocation":41},"https://ir.gitlab.com/","investor relations",{"text":327,"config":328},"Trust Center",{"href":329,"dataGaName":330,"dataGaLocation":41},"/security/","trust center",{"text":332,"config":333},"AI Transparency Center",{"href":334,"dataGaName":335,"dataGaLocation":41},"/ai-transparency-center/","ai transparency center",{"text":337,"config":338},"Newsletter",{"href":339,"dataGaName":340,"dataGaLocation":41},"/company/contact/#contact-forms","newsletter",{"text":342,"config":343},"Press",{"href":344,"dataGaName":345,"dataGaLocation":41},"/press/","press",{"text":347,"config":348,"lists":349},"Contact us",{"dataNavLevelOne":289},[350],{"items":351},[352,355,360],{"text":48,"config":353},{"href":50,"dataGaName":354,"dataGaLocation":41},"talk to sales",{"text":356,"config":357},"Support portal",{"href":358,"dataGaName":359,"dataGaLocation":41},"https://support.gitlab.com","support portal",{"text":361,"config":362},"Customer portal",{"href":363,"dataGaName":364,"dataGaLocation":41},"https://customers.gitlab.com/customers/sign_in/","customer portal",{"close":366,"login":367,"suggestions":374},"Close",{"text":368,"link":369},"To search repositories and projects, login to",{"text":370,"config":371},"gitlab.com",{"href":55,"dataGaName":372,"dataGaLocation":373},"search login","search",{"text":375,"default":376},"Suggestions",[377,379,383,385,389,393],{"text":70,"config":378},{"href":75,"dataGaName":70,"dataGaLocation":373},{"text":380,"config":381},"Code Suggestions (AI)",{"href":382,"dataGaName":380,"dataGaLocation":373},"/solutions/code-suggestions/",{"text":104,"config":384},{"href":106,"dataGaName":104,"dataGaLocation":373},{"text":386,"config":387},"GitLab on AWS",{"href":388,"dataGaName":386,"dataGaLocation":373},"/partners/technology-partners/aws/",{"text":390,"config":391},"GitLab on Google Cloud",{"href":392,"dataGaName":390,"dataGaLocation":373},"/partners/technology-partners/google-cloud-platform/",{"text":394,"config":395},"Why GitLab?",{"href":83,"dataGaName":394,"dataGaLocation":373},{"freeTrial":397,"mobileIcon":402,"desktopIcon":407,"secondaryButton":410},{"text":398,"config":399},"Start free trial",{"href":400,"dataGaName":46,"dataGaLocation":401},"https://gitlab.com/-/trials/new/","nav",{"altText":403,"config":404},"Gitlab Icon",{"src":405,"dataGaName":406,"dataGaLocation":401},"https://res.cloudinary.com/about-gitlab-com/image/upload/v1758203874/jypbw1jx72aexsoohd7x.svg","gitlab icon",{"altText":403,"config":408},{"src":409,"dataGaName":406,"dataGaLocation":401},"https://res.cloudinary.com/about-gitlab-com/image/upload/v1758203875/gs4c8p8opsgvflgkswz9.svg",{"text":411,"config":412},"Get Started",{"href":413,"dataGaName":414,"dataGaLocation":401},"https://gitlab.com/-/trial_registrations/new?glm_source=about.gitlab.com/get-started/","get started",{"freeTrial":416,"mobileIcon":421,"desktopIcon":423},{"text":417,"config":418},"Learn more about GitLab Duo",{"href":419,"dataGaName":420,"dataGaLocation":401},"/gitlab-duo/","gitlab duo",{"altText":403,"config":422},{"src":405,"dataGaName":406,"dataGaLocation":401},{"altText":403,"config":424},{"src":409,"dataGaName":406,"dataGaLocation":401},{"freeTrial":426,"mobileIcon":431,"desktopIcon":433},{"text":427,"config":428},"Back to pricing",{"href":183,"dataGaName":429,"dataGaLocation":401,"icon":430},"back to pricing","GoBack",{"altText":403,"config":432},{"src":405,"dataGaName":406,"dataGaLocation":401},{"altText":403,"config":434},{"src":409,"dataGaName":406,"dataGaLocation":401},{"title":436,"button":437,"config":442},"See how agentic AI transforms software delivery",{"text":438,"config":439},"Watch GitLab Transcend now",{"href":440,"dataGaName":441,"dataGaLocation":41},"/events/transcend/virtual/","transcend event",{"layout":443,"icon":444},"release","AiStar",{"data":446},{"text":447,"source":448,"edit":454,"contribute":459,"config":464,"items":469,"minimal":676},"Git is a trademark of Software Freedom Conservancy and our use of 'GitLab' is under license",{"text":449,"config":450},"View page source",{"href":451,"dataGaName":452,"dataGaLocation":453},"https://gitlab.com/gitlab-com/marketing/digital-experience/about-gitlab-com/","page source","footer",{"text":455,"config":456},"Edit this page",{"href":457,"dataGaName":458,"dataGaLocation":453},"https://gitlab.com/gitlab-com/marketing/digital-experience/about-gitlab-com/-/blob/main/content/","web ide",{"text":460,"config":461},"Please contribute",{"href":462,"dataGaName":463,"dataGaLocation":453},"https://gitlab.com/gitlab-com/marketing/digital-experience/about-gitlab-com/-/blob/main/CONTRIBUTING.md/","please contribute",{"twitter":465,"facebook":466,"youtube":467,"linkedin":468},"https://twitter.com/gitlab","https://www.facebook.com/gitlab","https://www.youtube.com/channel/UCnMGQ8QHMAnVIsI3xJrihhg","https://www.linkedin.com/company/gitlab-com",[470,517,571,615,642],{"title":181,"links":471,"subMenu":486},[472,476,481],{"text":473,"config":474},"View plans",{"href":183,"dataGaName":475,"dataGaLocation":453},"view plans",{"text":477,"config":478},"Why Premium?",{"href":479,"dataGaName":480,"dataGaLocation":453},"/pricing/premium/","why premium",{"text":482,"config":483},"Why Ultimate?",{"href":484,"dataGaName":485,"dataGaLocation":453},"/pricing/ultimate/","why ultimate",[487],{"title":488,"links":489},"Contact Us",[490,493,495,497,502,507,512],{"text":491,"config":492},"Contact sales",{"href":50,"dataGaName":51,"dataGaLocation":453},{"text":356,"config":494},{"href":358,"dataGaName":359,"dataGaLocation":453},{"text":361,"config":496},{"href":363,"dataGaName":364,"dataGaLocation":453},{"text":498,"config":499},"Status",{"href":500,"dataGaName":501,"dataGaLocation":453},"https://status.gitlab.com/","status",{"text":503,"config":504},"Terms of use",{"href":505,"dataGaName":506,"dataGaLocation":453},"/terms/","terms of use",{"text":508,"config":509},"Privacy statement",{"href":510,"dataGaName":511,"dataGaLocation":453},"/privacy/","privacy statement",{"text":513,"config":514},"Cookie preferences",{"dataGaName":515,"dataGaLocation":453,"id":516,"isOneTrustButton":26},"cookie preferences","ot-sdk-btn",{"title":86,"links":518,"subMenu":527},[519,523],{"text":520,"config":521},"DevSecOps platform",{"href":68,"dataGaName":522,"dataGaLocation":453},"devsecops platform",{"text":524,"config":525},"AI-Assisted Development",{"href":419,"dataGaName":526,"dataGaLocation":453},"ai-assisted development",[528],{"title":529,"links":530},"Topics",[531,536,541,546,551,556,561,566],{"text":532,"config":533},"CICD",{"href":534,"dataGaName":535,"dataGaLocation":453},"/topics/ci-cd/","cicd",{"text":537,"config":538},"GitOps",{"href":539,"dataGaName":540,"dataGaLocation":453},"/topics/gitops/","gitops",{"text":542,"config":543},"DevOps",{"href":544,"dataGaName":545,"dataGaLocation":453},"/topics/devops/","devops",{"text":547,"config":548},"Version Control",{"href":549,"dataGaName":550,"dataGaLocation":453},"/topics/version-control/","version control",{"text":552,"config":553},"DevSecOps",{"href":554,"dataGaName":555,"dataGaLocation":453},"/topics/devsecops/","devsecops",{"text":557,"config":558},"Cloud Native",{"href":559,"dataGaName":560,"dataGaLocation":453},"/topics/cloud-native/","cloud native",{"text":562,"config":563},"AI for Coding",{"href":564,"dataGaName":565,"dataGaLocation":453},"/topics/devops/ai-for-coding/","ai for coding",{"text":567,"config":568},"Agentic AI",{"href":569,"dataGaName":570,"dataGaLocation":453},"/topics/agentic-ai/","agentic ai",{"title":572,"links":573},"Solutions",[574,576,578,583,587,590,594,597,599,602,605,610],{"text":128,"config":575},{"href":123,"dataGaName":128,"dataGaLocation":453},{"text":117,"config":577},{"href":100,"dataGaName":101,"dataGaLocation":453},{"text":579,"config":580},"Agile development",{"href":581,"dataGaName":582,"dataGaLocation":453},"/solutions/agile-delivery/","agile delivery",{"text":584,"config":585},"SCM",{"href":113,"dataGaName":586,"dataGaLocation":453},"source code management",{"text":532,"config":588},{"href":106,"dataGaName":589,"dataGaLocation":453},"continuous integration & delivery",{"text":591,"config":592},"Value stream management",{"href":156,"dataGaName":593,"dataGaLocation":453},"value stream management",{"text":537,"config":595},{"href":596,"dataGaName":540,"dataGaLocation":453},"/solutions/gitops/",{"text":166,"config":598},{"href":168,"dataGaName":169,"dataGaLocation":453},{"text":600,"config":601},"Small business",{"href":173,"dataGaName":174,"dataGaLocation":453},{"text":603,"config":604},"Public sector",{"href":178,"dataGaName":179,"dataGaLocation":453},{"text":606,"config":607},"Education",{"href":608,"dataGaName":609,"dataGaLocation":453},"/solutions/education/","education",{"text":611,"config":612},"Financial services",{"href":613,"dataGaName":614,"dataGaLocation":453},"/solutions/finance/","financial services",{"title":186,"links":616},[617,619,621,623,626,628,630,632,634,636,638,640],{"text":198,"config":618},{"href":200,"dataGaName":201,"dataGaLocation":453},{"text":203,"config":620},{"href":205,"dataGaName":206,"dataGaLocation":453},{"text":208,"config":622},{"href":210,"dataGaName":211,"dataGaLocation":453},{"text":213,"config":624},{"href":215,"dataGaName":625,"dataGaLocation":453},"docs",{"text":236,"config":627},{"href":238,"dataGaName":239,"dataGaLocation":453},{"text":231,"config":629},{"href":233,"dataGaName":234,"dataGaLocation":453},{"text":241,"config":631},{"href":243,"dataGaName":244,"dataGaLocation":453},{"text":249,"config":633},{"href":251,"dataGaName":252,"dataGaLocation":453},{"text":254,"config":635},{"href":256,"dataGaName":257,"dataGaLocation":453},{"text":259,"config":637},{"href":261,"dataGaName":262,"dataGaLocation":453},{"text":264,"config":639},{"href":266,"dataGaName":267,"dataGaLocation":453},{"text":269,"config":641},{"href":271,"dataGaName":272,"dataGaLocation":453},{"title":287,"links":643},[644,646,648,650,652,654,656,660,665,667,669,671],{"text":294,"config":645},{"href":296,"dataGaName":289,"dataGaLocation":453},{"text":299,"config":647},{"href":301,"dataGaName":302,"dataGaLocation":453},{"text":307,"config":649},{"href":309,"dataGaName":310,"dataGaLocation":453},{"text":312,"config":651},{"href":314,"dataGaName":315,"dataGaLocation":453},{"text":317,"config":653},{"href":319,"dataGaName":320,"dataGaLocation":453},{"text":322,"config":655},{"href":324,"dataGaName":325,"dataGaLocation":453},{"text":657,"config":658},"Sustainability",{"href":659,"dataGaName":657,"dataGaLocation":453},"/sustainability/",{"text":661,"config":662},"Diversity, inclusion and belonging (DIB)",{"href":663,"dataGaName":664,"dataGaLocation":453},"/diversity-inclusion-belonging/","Diversity, inclusion and belonging",{"text":327,"config":666},{"href":329,"dataGaName":330,"dataGaLocation":453},{"text":337,"config":668},{"href":339,"dataGaName":340,"dataGaLocation":453},{"text":342,"config":670},{"href":344,"dataGaName":345,"dataGaLocation":453},{"text":672,"config":673},"Modern Slavery Transparency Statement",{"href":674,"dataGaName":675,"dataGaLocation":453},"https://handbook.gitlab.com/handbook/legal/modern-slavery-act-transparency-statement/","modern slavery transparency statement",{"items":677},[678,681,684],{"text":679,"config":680},"Terms",{"href":505,"dataGaName":506,"dataGaLocation":453},{"text":682,"config":683},"Cookies",{"dataGaName":515,"dataGaLocation":453,"id":516,"isOneTrustButton":26},{"text":685,"config":686},"Privacy",{"href":510,"dataGaName":511,"dataGaLocation":453},[688],{"id":689,"title":18,"body":8,"config":690,"content":692,"description":8,"extension":24,"meta":696,"navigation":26,"path":697,"seo":698,"stem":699,"__hash__":700},"blogAuthors/en-us/blog/authors/steve-azzopardi.yml",{"template":691},"BlogAuthor",{"name":18,"config":693},{"headshot":694,"ctfId":695},"","steveazz",{},"/en-us/blog/authors/steve-azzopardi",{},"en-us/blog/authors/steve-azzopardi","qnb6MwMS78l86ZK9JZJ-5qc-NcHCq2bdzgXn2tHRN40",[702,717,730],{"content":703,"config":715},{"title":704,"description":705,"authors":706,"heroImage":708,"date":709,"body":710,"category":9,"tags":711},"How to use GitLab Container Virtual Registry with Docker Hardened Images","Learn how to simplify container image management with this step-by-step guide.",[707],"Tim Rizzi","https://res.cloudinary.com/about-gitlab-com/image/upload/v1772111172/mwhgbjawn62kymfwrhle.png","2026-03-12","If you're a platform engineer, you've probably had this conversation:\n  \n*\"Security says we need to use hardened base images.\"*\n\n*\"Great, where do I configure credentials for yet another registry?\"*\n\n*\"Also, how do we make sure everyone actually uses them?\"*\n\nOr this one:\n\n*\"Why are our builds so slow?\"*\n\n*\"We're pulling the same 500MB image from Docker Hub in every single job.\"*\n\n*\"Can't we just cache these somewhere?\"*\n\nI've been working on [Container Virtual Registry](https://docs.gitlab.com/user/packages/virtual_registry/container/) at GitLab specifically to solve these problems. It's a pull-through cache that sits in front of your upstream registries — Docker Hub, dhi.io (Docker Hardened Images), MCR, and Quay — and gives your teams a single endpoint to pull from. Images get cached on the first pull. Subsequent pulls come from the cache. Your developers don't need to know or care which upstream a particular image came from.\n\nThis article shows you how to set up Container Virtual Registry, specifically with Docker Hardened Images in mind, since that's a combination that makes a lot of sense for teams concerned about security and not making their developers' lives harder.\n\n## What problem are we actually solving?\n\nThe Platform teams I usually talk to manage container images across three to five registries:\n\n* **Docker Hub** for most base images\n* **dhi.io** for Docker Hardened Images (security-conscious workloads)\n* **MCR** for .NET and Azure tooling\n* **Quay.io** for Red Hat ecosystem stuff\n* **Internal registries** for proprietary images\n\nEach one has its own:\n\n* Authentication mechanism\n* Network latency characteristics\n* Way of organizing image paths\n\nYour CI/CD configs end up littered with registry-specific logic. Credential management becomes a project unto itself. And every pipeline job pulls the same base images over the network, even though they haven't changed in weeks.\n\nContainer Virtual Registry consolidates this. One registry URL. One authentication flow (GitLab's). Cached images are served from GitLab's infrastructure rather than traversing the internet each time.\n\n## How it works\n\nThe model is straightforward:\n\n```text\nYour pipeline pulls:\n  gitlab.com/virtual_registries/container/1000016/python:3.13\n\nVirtual registry checks:\n  1. Do I have this cached? → Return it\n  2. No? → Fetch from upstream, cache it, return it\n\n```\n\nYou configure upstreams in priority order. When a pull request comes in, the virtual registry checks each upstream until it finds the image. The result gets cached for a configurable period (default 24 hours).\n\n```text\n┌─────────────────────────────────────────────────────────┐\n│                    CI/CD Pipeline                       │\n│                          │                              │\n│                          ▼                              │\n│   gitlab.com/virtual_registries/container/\u003Cid>/image   │\n└─────────────────────────────────────────────────────────┘\n                           │\n                           ▼\n┌─────────────────────────────────────────────────────────┐\n│            Container Virtual Registry                   │\n│                                                         │\n│  Upstream 1: Docker Hub ────────────────┐               │\n│  Upstream 2: dhi.io (Hardened) ────────┐│               │\n│  Upstream 3: MCR ─────────────────────┐││               │\n│  Upstream 4: Quay.io ────────────────┐│││               │\n│                                      ││││               │\n│                    ┌─────────────────┴┴┴┴──┐            │\n│                    │        Cache          │            │\n│                    │  (manifests + layers) │            │\n│                    └───────────────────────┘            │\n└─────────────────────────────────────────────────────────┘\n```\n\n## Why this matters for Docker Hardened Images\n\n[Docker Hardened Images](https://docs.docker.com/dhi/) are great because of the minimal attack surface, near-zero CVEs, proper software bills of materials (SBOMs), and SLSA provenance. If you're evaluating base images for security-sensitive workloads, they should be on your list.\n\nBut adopting them creates the same operational friction as any new registry:\n\n* **Credential distribution**: You need to get Docker credentials to every system that pulls images from dhi.io.\n* **CI/CD changes**: Every pipeline needs to be updated to authenticate with dhi.io.\n* **Developer friction**: People need to remember to use the hardened variants.\n* **Visibility gap**: It's difficulat to tell if teams are actually using hardened images vs. regular ones.\n\nVirtual registry addresses each of these:\n\n**Single credential**: Teams authenticate to GitLab. The virtual registry handles upstream authentication. You configure Docker credentials once, at the registry level, and they apply to all pulls.\n\n**No CI/CD changes per-team**: Point pipelines at your virtual registry. Done. The upstream configuration is centralized.\n\n**Gradual adoption**: Since images get cached with their full path, you can see in the cache what's being pulled. If someone's pulling `library/python:3.11` instead of the hardened variant, you'll know.\n\n**Audit trail**: The cache shows you exactly which images are in active use. Useful for compliance, useful for understanding what your fleet actually depends on.\n\n## Setting it up\n\nHere's a real setup using the Python client from this demo project.\n\n### Create the virtual registry\n\n```python\nfrom virtual_registry_client import VirtualRegistryClient\n\nclient = VirtualRegistryClient()\n\nregistry = client.create_virtual_registry(\n    group_id=\"785414\",  # Your top-level group ID\n    name=\"platform-images\",\n    description=\"Cached container images for platform teams\"\n)\n\nprint(f\"Registry ID: {registry['id']}\")\n# You'll need this ID for the pull URL\n```\n\n### Add Docker Hub as an upstream\n\nFor official images like Alpine, Python, etc.:\n\n```python\ndocker_upstream = client.create_upstream(\n    registry_id=registry['id'],\n    url=\"https://registry-1.docker.io\",\n    name=\"Docker Hub\",\n    cache_validity_hours=24\n)\n```\n\n### Add Docker Hardened Images (dhi.io)\n\nDocker Hardened Images are hosted on `dhi.io`, a separate registry that requires authentication:\n\n```python\ndhi_upstream = client.create_upstream(\n    registry_id=registry['id'],\n    url=\"https://dhi.io\",\n    name=\"Docker Hardened Images\",\n    username=\"your-docker-username\",\n    password=\"your-docker-access-token\",\n    cache_validity_hours=24\n)\n```\n\n### Add other upstreams\n\n```python\n# MCR for .NET teams\nclient.create_upstream(\n    registry_id=registry['id'],\n    url=\"https://mcr.microsoft.com\",\n    name=\"Microsoft Container Registry\",\n    cache_validity_hours=48\n)\n\n# Quay for Red Hat stuff\nclient.create_upstream(\n    registry_id=registry['id'],\n    url=\"https://quay.io\",\n    name=\"Quay.io\",\n    cache_validity_hours=24\n)\n```\n\n### Update your CI/CD\n\nHere's a `.gitlab-ci.yml` that pulls through the virtual registry:\n\n```yaml\nvariables:\n  VIRTUAL_REGISTRY_ID: \u003Cyour_virtual_registry_ID>\n\n  \nbuild:\n  image: docker:24\n  services:\n    - docker:24-dind\n  before_script:\n    # Authenticate to GitLab (which handles upstream auth for you)\n    - echo \"${CI_JOB_TOKEN}\" | docker login -u gitlab-ci-token --password-stdin gitlab.com\n  script:\n    # All of these go through your single virtual registry\n    \n    # Official Docker Hub images (use library/ prefix)\n    - docker pull gitlab.com/virtual_registries/container/${VIRTUAL_REGISTRY_ID}/library/alpine:latest\n    \n    # Docker Hardened Images from dhi.io (no prefix needed)\n    - docker pull gitlab.com/virtual_registries/container/${VIRTUAL_REGISTRY_ID}/python:3.13\n    \n    # .NET from MCR\n    - docker pull gitlab.com/virtual_registries/container/${VIRTUAL_REGISTRY_ID}/dotnet/sdk:8.0\n```\n\n### Image path formats\n\nDifferent registries use different path conventions:\n\n| Registry | Pull URL Example |\n|----------|------------------|\n| Docker Hub (official) | `.../library/python:3.11-slim` |\n| Docker Hardened Images (dhi.io) | `.../python:3.13` |\n| MCR | `.../dotnet/sdk:8.0` |\n| Quay.io | `.../prometheus/prometheus:latest` |\n\n### Verify it's working\n\nAfter some pulls, check your cache:\n\n```python\nupstreams = client.list_registry_upstreams(registry['id'])\nfor upstream in upstreams:\n    entries = client.list_cache_entries(upstream['id'])\n    print(f\"{upstream['name']}: {len(entries)} cached entries\")\n\n```\n\n## What the numbers look like\n\nI ran tests pulling images through the virtual registry:\n\n| Metric | Without Cache | With Warm Cache |\n|--------|---------------|-----------------|\n| Pull time (Alpine) | 10.3s | 4.2s |\n| Pull time (Python 3.13 DHI) | 11.6s | ~4s |\n| Network roundtrips to upstream | Every pull | Cache misses only |\n\n\n\n\nThe first pull is the same speed (it has to fetch from upstream). Every pull after that, for the cache validity period, comes straight from GitLab's storage. No network hop to Docker Hub, dhi.io, MCR, or wherever the image lives.\n\nFor a team running hundreds of pipeline jobs per day, that's hours of cumulative build time saved.\n\n## Practical considerations\nHere are some considerations to keep in mind:\n\n### Cache validity\n\n24 hours is the default. For security-sensitive images where you want patches quickly, consider 12 hours or less:\n\n```python\nclient.create_upstream(\n    registry_id=registry['id'],\n    url=\"https://dhi.io\",\n    name=\"Docker Hardened Images\",\n    username=\"your-username\",\n    password=\"your-token\",\n    cache_validity_hours=12\n)\n```\n\nFor stable, infrequently-updated images (like specific version tags), longer validity is fine.\n\n### Upstream priority\n\nUpstreams are checked in order. If you have images with the same name on different registries, the first matching upstream wins.\n\n### Limits\n\n* Maximum of 20 virtual registries per group\n* Maximum of 20 upstreams per virtual registry\n\n## Configuration via UI\n\nYou can also configure virtual registries and upstreams directly from the GitLab UI—no API calls required. Navigate to your group's **Settings > Packages and registries > Virtual Registry** to:\n\n* Create and manage virtual registries\n* Add, edit, and reorder upstream registries\n* View and manage the cache\n* Monitor which images are being pulled\n\n## What's next\n\nWe're actively developing:\n\n* **Allow/deny lists**: Use regex to control which images can be pulled from specific upstreams.\n\nThis is beta software. It works, people are using it in production, but we're still iterating based on feedback.\n\n## Share your feedback\n\nIf you're a platform engineer dealing with container registry sprawl, I'd like to understand your setup:\n\n* How many upstream registries are you managing?\n* What's your biggest pain point with the current state?\n* Would something like this help, and if not, what's missing?\n\nPlease share your experiences in the [Container Virtual Registry feedback issue](https://gitlab.com/gitlab-org/gitlab/-/work_items/589630).\n## Related resources\n- [New GitLab metrics and registry features help reduce CI/CD bottlenecks](https://about.gitlab.com/blog/new-gitlab-metrics-and-registry-features-help-reduce-ci-cd-bottlenecks/#container-virtual-registry)\n- [Container Virtual Registry documentation](https://docs.gitlab.com/user/packages/virtual_registry/container/)\n- [Container Virtual Registry API](https://docs.gitlab.com/api/container_virtual_registries/)",[712,713,714],"tutorial","product","features",{"featured":12,"template":13,"slug":716},"using-gitlab-container-virtual-registry-with-docker-hardened-images",{"content":718,"config":728},{"title":719,"description":720,"authors":721,"heroImage":723,"date":724,"category":9,"tags":725,"body":727},"How IIT Bombay students are coding the future with GitLab","At GitLab, we often talk about how software accelerates innovation. But sometimes, you have to step away from the Zoom calls and stand in a crowded university hall to remember why we do this.",[722],"Nick Veenhof","https://res.cloudinary.com/about-gitlab-com/image/upload/v1750099013/Blog/Hero%20Images/Blog/Hero%20Images/blog-image-template-1800x945%20%2814%29_6VTUA8mUhOZNDaRVNPeKwl_1750099012960.png","2026-01-08",[257,609,726],"open source","The GitLab team recently had the privilege of judging the **iHack Hackathon** at **IIT Bombay's E-Summit**. The energy was electric, the coffee was flowing, and the talent was undeniable. But what struck us most wasn't just the code — it was the sheer determination of students to solve real-world problems, often overcoming significant logistical and financial hurdles to simply be in the room.\n\n\nThrough our [GitLab for Education program](https://about.gitlab.com/solutions/education/), we aim to empower the next generation of developers with tools and opportunity. Here is a look at what the students built, and how they used GitLab to bridge the gap between idea and reality.\n\n## The challenge: Build faster, build securely\n\nThe premise for the GitLab track of the hackathon was simple: Don't just show us a product; show us how you built it. We wanted to see how students utilized GitLab's platform — from Issue Boards to CI/CD pipelines — to accelerate the development lifecycle.\n\nThe results were inspiring.\n\n## The winners\n\n### 1st place: Team Decode — Democratizing Scientific Research\n\n**Project:** FIRE (Fast Integrated Research Environment)\n\nTeam Decode took home the top prize with a solution that warms a developer's heart: a local-first, blazing-fast data processing tool built with [Rust](https://about.gitlab.com/blog/secure-rust-development-with-gitlab/) and Tauri. They identified a massive pain point for data science students: existing tools are fragmented, slow, and expensive.\n\nTheir solution, FIRE, allows researchers to visualize complex formats (like NetCDF) instantly. What impressed the judges most was their \"hacker\" ethos. They didn't just build a tool; they built it to be open and accessible.\n\n**How they used GitLab:** Since the team lived far apart, asynchronous communication was key. They utilized **GitLab Issue Boards** and **Milestones** to track progress and integrated their repo with Telegram to get real-time push notifications. As one team member noted, \"Coordinating all these technologies was really difficult, and what helped us was GitLab... the Issue Board really helped us track who was doing what.\"\n\n![Team Decode](https://res.cloudinary.com/about-gitlab-com/image/upload/v1767380253/epqazj1jc5c7zkgqun9h.jpg)\n\n### 2nd place: Team BichdeHueDost — Reuniting to Solve Payments\n\n**Project:** SemiPay (RFID Cashless Payment for Schools)\n\nThe team name, BichdeHueDost, translates to \"Friends who have been set apart.\" It's a fitting name for a group of friends who went to different colleges but reunited to build this project. They tackled a unique problem: handling cash in schools for young children. Their solution used RFID cards backed by a blockchain ledger to ensure secure, cashless transactions for students.\n\n**How they used GitLab:** They utilized [GitLab CI/CD](https://about.gitlab.com/topics/ci-cd/) to automate the build process for their Flutter application (APK), ensuring that every commit resulted in a testable artifact. This allowed them to iterate quickly despite the \"flaky\" nature of cross-platform mobile development.\n\n![Team BichdeHueDost](https://res.cloudinary.com/about-gitlab-com/image/upload/v1767380253/pkukrjgx2miukb6nrj5g.jpg)\n\n### 3rd place: Team ZenYukti — Agentic Repository Intelligence\n\n**Project:** RepoInsight AI (AI-powered, GitLab-native intelligence platform)\n\nTeam ZenYukti impressed us with a solution that tackles a universal developer pain point: understanding unfamiliar codebases. What stood out to the judges was the tool's practical approach to onboarding and code comprehension: RepoInsight-AI automatically generates documentation, visualizes repository structure, and even helps identify bugs, all while maintaining context about the entire codebase.\n\n**How they used GitLab:** The team built a comprehensive CI/CD pipeline that showcased GitLab's security and DevOps capabilities. They integrated [GitLab's Security Templates](https://gitlab.com/gitlab-org/gitlab/-/tree/master/lib/gitlab/ci/templates/Security) (SAST, Dependency Scanning, and Secret Detection), and utilized [GitLab Container Registry](https://docs.gitlab.com/user/packages/container_registry/) to manage their Docker images for backend and frontend components. They created an AI auto-review bot that runs on merge requests, demonstrating an \"agentic workflow\" where AI assists in the development process itself.\n\n![Team ZenYukti](https://res.cloudinary.com/about-gitlab-com/image/upload/v1767380253/ymlzqoruv5al1secatba.jpg)\n\n## Beyond the code: A lesson in inclusion\n\nWhile the code was impressive, the most powerful moment of the event happened away from the keyboard.\n\nDuring the feedback session, we learned about the journey Team ZenYukti took to get to Mumbai. They traveled over 24 hours, covering nearly 1,800 kilometers. Because flights were too expensive and trains were booked, they traveled in the \"General Coach,\" a non-reserved, severely overcrowded carriage.\n\nAs one student described it:\n\n*\"You cannot even imagine something like this... there are no seats... people sit on the top of the train. This is what we have endured.\"*\n\nThis hit home. [Diversity, Inclusion, and Belonging](https://handbook.gitlab.com/handbook/company/culture/inclusion/) are core values at GitLab. We realized that for these students, the barrier to entry wasn't intellect or skill, it was access.\n\nIn that moment, we decided to break that barrier. We committed to reimbursing the travel expenses for the participants who struggled to get there. It's a small step, but it underlines a massive truth: **talent is distributed equally, but opportunity is not.**\n\n![hackathon class together](https://res.cloudinary.com/about-gitlab-com/image/upload/v1767380252/o5aqmboquz8ehusxvgom.jpg)\n\n### The future is bright (and automated)\n\nWe also saw incredible potential in teams like Prometheus, who attempted to build an autonomous patch remediation tool (DevGuardian), and Team Arrakis, who built a voice-first job portal for blue-collar workers using [GitLab Duo](https://about.gitlab.com/gitlab-duo/) to troubleshoot their pipelines.\n\nTo all the students who participated: You are the future. Through [GitLab for Education](https://about.gitlab.com/solutions/education/), we are committed to providing you with the top-tier tools (like GitLab Ultimate) you need to learn, collaborate, and change the world — whether you are coding from a dorm room, a lab, or a train carriage. **Keep shipping.**\n\n> :bulb: Learn more about the [GitLab for Education program](https://about.gitlab.com/solutions/education/).\n",{"slug":729,"featured":12,"template":13},"how-iit-bombay-students-code-future-with-gitlab",{"content":731,"config":739},{"title":732,"description":733,"authors":734,"heroImage":735,"date":736,"category":9,"tags":737,"body":738},"Artois University elevates research and curriculum with GitLab Ultimate for Education","Artois University's CRIL leveraged the GitLab for Education program to gain free access to Ultimate, transforming advanced research and computer science curricula.",[722],"https://res.cloudinary.com/about-gitlab-com/image/upload/v1750099203/Blog/Hero%20Images/Blog/Hero%20Images/blog-image-template-1800x945%20%2820%29_2bJGC5ZP3WheoqzlLT05C5_1750099203484.png","2025-12-10",[609,257,713],"Leading academic institutions face a critical challenge: how to provide thousands of students and researchers with industry-standard, **full-featured DevSecOps tools** without compromising institutional control. Many start with basic version control, but the modern curriculum demands integrated capabilities for planning, security, and advanced CI/CD.\n\nThe **GitLab for Education program** is designed to solve this by providing access to **GitLab Ultimate** for qualifying institutions, allowing them to scale their operations and elevate their academic offerings. \n\nThis article showcases a powerful success story from the **Centre de Recherche en Informatique de Lens (CRIL)**, a joint laboratory of **Artois University** and CNRS in France. After years of relying solely on GitLab Community Edition (CE), the university's move to GitLab Ultimate through the GitLab for Education program immediately unlocked advanced capabilities, transforming their teaching, research, and contribution workflows virtually overnight. This story demonstrates why GitLab Ultimate is essential for institutions seeking to deliver advanced computer science and research curricula.\n\n## GitLab Ultimate unlocked: Managing scale and driving academic value\n\n**Artois University's** self-managed GitLab instance is a large-scale operation, supporting nearly **3,000 users** across approximately **19,000 projects**, primarily serving computer science students and researchers. While GitLab Community Edition was robust, the upgrade to GitLab Ultimate provided the sophisticated tooling necessary for managing this scale and facilitating advanced university-level work.\n\n***\"We can see the difference,\" says Daniel Le Berre, head of research at CRIL and the instance maintainer. \"It's a completely different product. Each week reveals new features that directly enhance our productivity and teaching.\"***\n\nThe institution joined the GitLab for Education program specifically because it covers both **instructional and non-commercial research use cases** and offers full access to Ultimate's features, removing significant cost barriers.\n\n### Key GitLab Ultimate benefits for students and researchers\n\n* **Advanced project management at scale:** Master's students now benefit from **GitLab Ultimate's project planning features**. This enables them to structure, track, and manage complex, long-term research projects using professional methodologies like portfolio management and advanced issue tracking that seamlessly roll up across their thousands of projects.\n\n* **Enhanced visibility:** Features like improved dashboards and code previews directly in Markdown files dramatically streamline tracking and documentation review, reducing administrative friction for both instructors and students managing large project loads.\n\n## Comprehensive curriculum: From concepts to continuous delivery\n\nGitLab Ultimate is deeply integrated into the computer science curriculum, moving students beyond simple `git` commands to practical **DevSecOps implementation**.\n\n* **Git fundamentals:** Students begin by visualizing concepts using open-source tools to master Git concepts.\n\n* **Full CI/CD implementation:** Students use GitLab CI for rigorous **Test-Driven Development (TDD)** in their software projects. They learn to build, test, and perform quality assurance using unit and integration testing pipelines—core competency made seamless by the integrated platform.\n\n* **DevSecOps for research and documentation:** The university teaches students that DevSecOps principles are vital for all collaborative work. Inspired by earlier work in Delft, students manage and produce critical research documentation (PDFs from Markdown files) using GitLab, incorporating quality checks like linters and spell checks directly in the CI pipeline. This ensures high-quality, reproducible research output.\n\n* **Future-proofing security skills:** The GitLab Ultimate platform immediately positions the institution to incorporate advanced DevSecOps features like SAST and DAST scanning as their research and development code projects grow, ensuring students are prepared for industry security standards.\n\n## Accelerating open source contributions with GitLab Duo\n\nAccess to the full GitLab platform, including our AI capabilities, has empowered students to make impactful contributions to the wider open source community faster than ever before.\n\nTwo Master's students recently completed direct contributions to the GitLab product, adding the **ORCID identifier** into user profiles. Working on GitLab.com, they leveraged **GitLab Duo's AI chat and code suggestions** to navigate the codebase efficiently.\n\n***\"This would not have been possible without GitLab Duo,\" Daniel Le Berre notes. \"The AI features helped students, who might have lacked deep codebase knowledge, deliver meaningful contributions in just two weeks.\"***\n\nThis demonstrates how providing students with cutting-edge tools **accelerates their learning and impact**, allowing them to translate classroom knowledge into real-world contributions immediately.\n\n## Empowering open research and institutional control\n\nThe stability of the self-managed instance at Artois University is key to its success. This model guarantees **institutional control and stability** — a critical factor for long-term research preservation.\n\nThe institution's expertise in this area was recently highlighted in a major 2024 study led by CRIL, titled: \"[Higher Education and Research Forges in France - Definition, uses, limitations encountered and needs analysis](https://hal.science/hal-04208924v4)\" ([Project on GitLab](https://gitlab.in2p3.fr/coso-college-codes-sources-et-logiciels/forges-esr-en)). The research found that the vast majority of public forges in French Higher Education and Research relied on **GitLab**. This finding underscores the consensus among academic leaders that self-hosted solutions are essential for **data control and longevity**, especially when compared to relying on external, commercial forges.\n\n## Unlock GitLab Ultimate for your institution today\n\nThe success story of **Artois University's CRIL** proves the transformative power of the GitLab for Education program. By providing **free access to GitLab Ultimate**, we enable large-scale institutions to:\n\n1.  **Deliver a modern, integrated DevSecOps curriculum.**\n\n2.  **Support advanced, collaborative research projects with Ultimate planning features.**\n\n3.  **Empower students to make AI-assisted open source contributions.**\n\n4.  **Maintain institutional control and data longevity.**\n\nIf your academic institution is ready to equip its students and researchers with the complete DevSecOps platform and its most advanced features, we invite you to join the program.\n\nThe program provides **free access to GitLab Ultimate** for qualifying instructional and non-commercial research use cases.\n\n**Apply now [online](https://about.gitlab.com/solutions/education/join/).**\n",{"slug":740,"featured":26,"template":13},"artois-university-elevates-curriculum-with-gitlab-ultimate-for-education",{"promotions":742},[743,757,768],{"id":744,"categories":745,"header":747,"text":748,"button":749,"image":754},"ai-modernization",[746],"ai-ml","Is AI achieving its promise at scale?","Quiz will take 5 minutes or less",{"text":750,"config":751},"Get your AI maturity score",{"href":752,"dataGaName":753,"dataGaLocation":239},"/assessments/ai-modernization-assessment/","modernization assessment",{"config":755},{"src":756},"https://res.cloudinary.com/about-gitlab-com/image/upload/v1772138786/qix0m7kwnd8x2fh1zq49.png",{"id":758,"categories":759,"header":760,"text":748,"button":761,"image":765},"devops-modernization",[713,555],"Are you just managing tools or shipping innovation?",{"text":762,"config":763},"Get your DevOps maturity score",{"href":764,"dataGaName":753,"dataGaLocation":239},"/assessments/devops-modernization-assessment/",{"config":766},{"src":767},"https://res.cloudinary.com/about-gitlab-com/image/upload/v1772138785/eg818fmakweyuznttgid.png",{"id":769,"categories":770,"header":772,"text":748,"button":773,"image":777},"security-modernization",[771],"security","Are you trading speed for security?",{"text":774,"config":775},"Get your security maturity score",{"href":776,"dataGaName":753,"dataGaLocation":239},"/assessments/security-modernization-assessment/",{"config":778},{"src":779},"https://res.cloudinary.com/about-gitlab-com/image/upload/v1772138786/p4pbqd9nnjejg5ds6mdk.png",{"header":781,"blurb":782,"button":783,"secondaryButton":788},"Start building faster today","See what your team can do with the intelligent orchestration platform for DevSecOps.\n",{"text":784,"config":785},"Get your free trial",{"href":786,"dataGaName":46,"dataGaLocation":787},"https://gitlab.com/-/trial_registrations/new?glm_content=default-saas-trial&glm_source=about.gitlab.com/","feature",{"text":491,"config":789},{"href":50,"dataGaName":51,"dataGaLocation":787},1773350813813]