- Docs
- Resources
- Deploys & Repos
- Troubleshooting
Troubleshooting
Find the message or symptom you see. Each entry says why it happens, what to check and how to fix it. Messages from the Git connection and the webhook come from the server in Turkish; each entry gives the meaning.
In short
On this page you can
- Fix repository form errors such as an invalid service name or a missing token.
- Read Test connection failures and what each one means.
- Fix webhook errors your CI log shows: 401, missing environment, bad SHA, unknown status.
- Find out why a deploy stays Not tested or has no commits.
How to use it
- Copy the exact message from the screen or the CI log.
- Find it below (use the browser’s search for a word of the message).
- Check the listed points in order.
- Apply the fix, then deploy again or click Refresh.
Repository form#
“Repository URL is required.”Repository
- Fix
- Paste the address of the repository in Repository URL. See URL formats.
“Service name: lowercase letters, digits, - or _ (e.g. web, api).”Repository
- Why
- The service name is empty, longer than 40 characters, starts with
-or_, or contains spaces or other characters. - Fix
- Use a short key such as
web,apiorauth-service. Upper case is turned to lower case for you.
“Bu projede 'web' servis adiyla baska bir repo var.” (another repository in this project uses this service name)Repository
- Why
- Service names are unique in a project, so deploys of two repositories never mix in one row.
- Check
- The Code repositories list: the chip
Service: webshows which repository already uses it. - Fix
- Pick another name, or edit the existing repository with Settings instead of adding a second one.
“Enter an access token or choose "Deploy notification only".”Repository
- Why
- A new repository with Connect with token needs a token.
- Fix
- Paste a read-only token (which one), or switch to Deploy notification only. You can add a token later.
“Repo adresi {sahip}/{repo} icermeli.”, “Azure DevOps adresi .../{proje}/_git/{repo} bicminde olmali.” or “Repo adresi http(s) olmali.”Repository
- Why
- The address must contain owner and repository; Azure DevOps addresses must contain
/{project}/_git/{repo}; onlyhttp,httpsandgit@host:owner/repoare accepted. Related: “Repo adresinde sunucu adi yok.” (no host name) and “Repo adresi gecersiz.” (the address cannot be read). - Fix
- Copy the address from the repository’s page in your browser, as in URL formats. If the server is not one of the listed providers, choose Other.
Add environment does nothingRepository
- Why
- The name is already in the list, or it has characters other than letters, digits,
-and_, or it is longer than 32 characters. Invalid names are ignored without a message. - Fix
- Type a name such as
STAGINGorUAT-2. A repository can map 12 environments; more gives “En fazla 12 ortam eslenebilir.” (at most 12 environments).
“Could not load.” or “Could not save.”Page
- Why
- The request failed without a more specific reason, usually because the session expired or the connection dropped.
- Fix
- Click Refresh. If it repeats, log in again. If you are a User without access to this project, ask an admin for project access.
“Select a project from the header to manage its agents.”Page
- Fix
- No project is selected. Pick one in the header; repositories belong to a project.
Test connection and commits#
These messages appear under Test connection, or in an amber box in the deploy detail when commits could not be read.
Test connection is greyed outConnection
- Fix
- Fill Repository URL and type a token. When you edit a repository with a saved token, the button works with the saved one.
“Token reddedildi (HTTP 401). Yetkiyi ve suresini kontrol edin.” (token rejected; check its rights and expiry)Connection
- Why
- The provider refused the token (HTTP 401 or 403): it expired, was revoked, lacks read rights, or belongs to another provider.
- Check
- The token type matches the provider table.
- A GitHub fine-grained token includes this repository.
- Bitbucket app passwords are written as
username:app-password.
- Fix
- Create a new read-only token, paste it in Settings and click Test connection.
“Repo ya da commit bulunamadi (HTTP 404)…” (repository or commit not found)Connection
- Why
- The URL is wrong, the token cannot see the repository (providers answer 404 instead of 403), or a commit CI reported was never pushed to the provider.
- Fix
- Check the URL and the token’s scope. In the deploy detail, check that CI sends the SHA of a pushed commit, then use Read commits again.
“Git sunucusuna ulasilamadi…”, “Git sunucusu zamaninda yanit vermedi.” or “Sunucu adi cozumlenemedi: …” (cannot reach, no answer in time, host name unknown)Connection
- Why
- The SmartestQA server cannot reach your Git server: it is inside your network, a firewall blocks it, or the name only resolves internally.
- Fix
- Allow SmartestQA to reach the Git server’s API, or use Deploy notification only: deploys are still recorded, only commits are missing.
“Bu adrese istek atilamaz: …” (requests to this address are not allowed)Connection
- Why
- The address points to the server itself (loopback such as
localhost) or to a link-local address. These are blocked for security. - Fix
- Use the Git server’s real host name or private network address in the URL or the API base URL.
“Git sunucusu yonlendirme dondu (HTTP 3xx). API adresini kontrol edin.” or “Git sunucusunun yaniti JSON degil. API adresini kontrol edin.” (redirect, or not a JSON answer; check the API address)Connection
- Why
- The address answers with a login page or a redirect instead of the API. SmartestQA does not follow redirects.
- Fix
- Set API base URL (self-hosted) to the API root, for example
https://git.acme.local/api/v4for GitLab. Usehttps://if the server redirects fromhttp://.
“Git sunucusu hata dondu (HTTP 500).” (the Git server returned an error) or “Connection test failed.”Connection
- Fix
- The provider had an error, or the request failed. Try again in a few minutes; check the provider’s status page.
“Commit list is shown when the repository is connected with a token.”Commits
- Why
- The repository uses Deploy notification only or the provider is Other.
- Fix
- Open the repository’s Settings, choose Connect with token, paste a token and save. Open the deploy again; older deploys get their commits when you open them. Provider Other never reads commits.
“Commit listesi okunamadi.” or “Kayitli commit listesi okunamadi.” (the commit list could not be read)Commits
- Fix
- Click Read commits again in the deploy detail. If it repeats, run Test connection in the repository’s Settings.
Webhook errors from CI#
These appear in your CI log, in the message of the response. The ready-made steps end with || true, so the deploy itself still succeeds; only the record is missing.
HTTP 401 “Gecersiz deploy token'i.” (invalid deploy token)Webhook
- Why
- The token is missing, mistyped, was replaced with Generate new token, or its repository was deleted.
- Check
- The header is
X-Deploy-Token(orAuthorization: Bearer sdt_…). - The CI secret starts with
sdt_and has no space or line break. - The last four characters match “Token ending in …” in CI code & deploy token.
- The header is
- Fix
- Generate a new token and update the CI secret.
HTTP 400 “environment gerekli (DEV, TEST, PRP ...).” (environment required)Webhook
- Why
environmentis empty or has characters other than letters, digits,-and_.- Fix
- Send
"environment":"TEST"(or DEV, PRP, or your own name). See the fields.
HTTP 400 “commitSha 7-64 karakterlik hex olmali.” (commitSha must be 7–64 hex characters)Webhook
- Why
- The SHA is empty (the CI variable was not set), shorter than 7 characters, or is a branch or tag name.
- Fix
- Use the commit variable of your CI:
${{ github.sha }},$CI_COMMIT_SHA,${env.GIT_COMMIT}.
HTTP 400 “Bilinmeyen status: … (success | failure | cancelled | started)” (unknown status)Webhook
- Fix
- Send one of the accepted values, or leave
statusempty for success. Jenkins’SUCCESS/FAILUREwork;UNSTABLEdoes not, so map it first as the ready-made Jenkins step does.
HTTP 400 “deployedAt ISO-8601 olmali (ornek 2026-10-06T21:59:00Z).” (deployedAt must be ISO-8601)Webhook
- Fix
- Send
2026-11-06T18:59:00Z, a time with offset (+03:00), Unix seconds or milliseconds, or leave it out to use the arrival time.
HTTP 400 “Govde bos.” (empty body)Webhook
- Fix
- Send the JSON body with
Content-Type: application/json. In shells, check that quotes around the-dvalue survived.
CI ran but nothing appears; the chip still says No deploy received yetWebhook
- Check
- The CI log of the notify step: an error message above means one of the entries on this page.
- The step ran at all: it must be after the deploy and run always (
if: always(),when: always,post { always }). - The URL is the Webhook URL of this SmartestQA server, with
POST. - You clicked Refresh: the page does not update by itself.
- Fix
- Run the curl tab’s command from the CI machine to see the answer directly.
Test state#
A deploy stays Not tested although a pipeline ranOverview
- Why
- The run is outside the deploy’s window or in another environment.
- Check
- Fix
- Run the pipeline in the right environment after the deploy. To make this automatic, start a pipeline from CI after the notify step, or let an Autonomous Agent react to deploys.
No pipeline environmentOverview
- Why
- The deploy went to an environment of your own, such as STAGING. Pipelines run only in DEV, TEST or PRP, so nothing can be matched.
- Fix
- If STAGING is where your TEST pipelines run, send
"environment":"TEST"from CI. Otherwise the cell shows the version only; that is expected. See the effect on the KPI cards.
“Last deploy failed” under a green chipOverview
- Why
- The newest deploy failed or was cancelled, so the previous successful version still runs. The chip describes that version. See Cell states.
- Fix
- Open Deploy history and the failed row, then Open CI run to see why it failed.
A repository has no row in the grid, or deploys are missing from itOverview
- Why
- The repository was deleted (its records stay in the history only), or CI uses the token of another repository, so the deploys land in that repository’s row.
- Fix
- Add the repository again and put its new token in CI. Check that each CI job uses the token of its own repository.
The deploy token is not shown any moreToken
- Why
- SmartestQA shows a token only once. It keeps a fingerprint, not the token.
- Fix
- If CI still works, nothing to do. If you need the value, generate a new token and update CI.
Didn’t find your answer?Write to support with your project name and what you see on screen.
info@smartestqa.ai