Repository navigation
ci: repair base-branch CRDA and Labeler workflows - #18
Conversation
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
|
ⓘ Qodo reviews are paused because your trial has ended. Ask your workspace admin to add credits to resume reviews. Manage billing |
|
You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard. |
Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> Signed-off-by: zskbot <241179063+zskbot@users.noreply.github.com>
Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> Signed-off-by: zskbot <241179063+zskbot@users.noreply.github.com>
🚀 Deploying Preview to Cloudflare 🚀Preview Deployments by commit
|
|
On this page LimitsRequest a limit increaseYou can raise many of the limits on this page. The process depends on your plan:
General limitsTo prevent abuse of our platform, we apply the following limits to all accounts.
Hobby teams get a monthly allotment of each billable resource. Pro teams get a monthly credit that applies across billable resources, and Vercel charges for additional usage on demand once you use that credit. For the allotments and the on-demand rates, see typical monthly usage guidelines. Pro trial limitsSee the Pro trial limitations section for information on the limits that apply to Pro trials. Routes created per deploymentThe limit of "Routes created per Deployment" encapsulates several options that can be configured on Vercel:
Note that most frameworks will create Routes automatically for you. For example, Next.js will create a set of Routes corresponding to your use of dynamic routes, redirects, rewrites and custom headers. Build time per deploymentThe maximum duration of the Build Step is 45 minutes. When the limit is reached, the Build Step will be interrupted and the Deployment will fail. Build container resourcesEvery build container has a fixed amount of resources available to it. You can find the resources available for each build machine type here. For more information on troubleshooting these, see Build container resources. Static file uploadsWhen using the CLI to deploy, the maximum size of the source files that can be uploaded is limited to 100 MB for Hobby and 1 GB for Pro. If the size of the source files exceeds this limit, the deployment will fail. Build cache maximum sizeThe maximum size of the Build's cache depends on the build machine type:
Elastic build machines use the limit of the machine type assigned to each build. The cache is retained for one month and applies at the level of each Build cache key. Observability PlusSee the Observability Plus limitations section for the limits that apply to Observability Plus, which includes Observability Query. Monitoring is deprecated and replaced by Observability Plus. If your team still has access to Monitoring, see Monitoring limits and pricing. LogsThere are two types of logs: build logs and runtime logs. Both have different behaviors when storing logs. Build logs are stored indefinitely for each deployment. Runtime logs are stored for 1 hour on Hobby, 1 day on Pro, and for 3 days on Enterprise accounts. To learn more about these log limits, read here. Environment variablesThe maximum number of Environment Variables per environment per Project is The total size of your Environment Variables, names and values, is limited to 64KB for projects using Node.js, Python, Ruby, Go, Java, and .NET runtimes. This limit is the total allowed for each deployment, and is also the maximum size of any single Environment Variable. For more information, see the Environment Variables documentation. If you are using System Environment Variables, the framework-specific ones (i.e. those prefixed by the framework name) are exposed only during the Build Step, but not at runtime. However, the non-framework-specific ones are exposed at runtime. Only the Environment Variables that are exposed at runtime are counted towards the size limit. Domains
FilesThe maximum number of files that can be uploaded when creating a CLI Deployment is Although there is no upper limit for output files created during a build, you can expect longer build times as a result of having many thousands of output files (100,000 or more, for example). If the build time exceeds 45 minutes then the build will fail. We recommend using Incremental Static Regeneration (ISR) to help reduce build time. Using ISR will allow you pre-render a subset of the total number of pages at build time, giving you faster builds and the ability to generate pages on-demand. Proxied request timeoutThe amount of time (in seconds) that a proxied request ( Web AnalyticsSee the Limits and Pricing section for more details about the limits of Vercel Web Analytics. Speed InsightsSee the Limits and Pricing doc for more details about the limits of the Speed Insights feature on Vercel. Cron JobsSee the Cron Jobs limits section for more information about the limits of Cron Jobs on Vercel. Vercel Workflows limitsVercel Workflows limits the size and throughput of each workflow run, including events and steps per run, payload size, stored data per run, and stream throughput. It also rate limits the requests your team's workflows make to the Workflows backend, such as creating events, writing to streams, and reading run data. Requests over the rate limit are retried automatically with backoff, so hitting it slows your workflows down but does not cause them to fail. See Workflow limits and Workflow rate limits for the values on each plan. Vercel FunctionsVercel Functions run on Fluid compute, which is enabled by default for new projects. Its key limits per plan:
For the full reference, see Vercel Functions limitations. Connecting a project to a Git repositoryVercel does not support connecting a project on your Hobby team to Git repositories owned by Git organizations. You can either switch to an existing Team or create a new one. The same limitation applies in the Project creation flow when importing an existing Git repository or when cloning a Vercel template to a new Git repository as part of your Git organization. Reserved variablesSee the Reserved Environment Variables docs for more information. Rate limitsRate limits are hard limits that apply to the platform when performing actions that require a response from our API. The rate limits table consists of the following five columns:
Rate limit examplesBelow are six examples that provide further information on how rate limits work. Domain deletionYou are able to delete up to Team deletionYou are able to delete up to Username changeYou are able to change your username up to Builds per hour (Hobby)You are able to build Using Next.js or any similar framework to build your deployment is classed as a build. Each Vercel Function is also classed as a build. Hosting static files such as an index.html file is not classed as a build. Deployments per day (Hobby)You are able to deploy Deploy hook triggersYou can trigger deploy hooks up to The following table lists all API rate limits that apply when using the Vercel REST API. These limits apply to actions such as deployments, domain management, team operations, and more.
Dynamic quotasDynamic quotas cap how fast your provisioning rate for a resource can grow. They do not cap the total number of actions in a fixed window. Each dynamic quota has an allowed rate, measured per minute, that moves between a minimum and a maximum. The unit depends on the resource and is named in the limit's description.
Your allowance is one minute's worth of your current allowed rate, so you can spend a full minute of capacity in one burst and then wait for it to refill. For example, take a limit with a minimum rate of 100 units per minute, a ramp of 200 units per minute, and a maximum rate of 1,000 units per minute. After 10 minutes with no requests you start at 100 units per minute, which gives you an allowance of 100 units. Keep consuming at or above your allowed rate and it climbs to 300 units per minute after one minute, 500 after two minutes, and reaches the 1,000 maximum after four and a half minutes of sustained usage. Drop below your allowed rate and it falls back toward 100 units per minute at the same slope. Once your recent usage has drawn down your allowance, the API responds with a
A single request can be larger than your full allowance. When your allowance is full, that request still goes through, but it drains your allowance below zero. Later requests get a The dynamic quotas table consists of the following five columns:
|
|
On this page TypeScript SDK Reference
For Python, see the Python SDK Reference. Installterminal The SDK has one runtime dependency, AuthenticationIn a Vercel deployment, the SDK reads To override the token explicitly, pass app/lib/explicit-token.ts See Authentication for the full caller-to-Vercel-Connect story. getTokenReturns the access token string. Use when you just need the token to put in an Rate limit: 200 requests per minute per team. signature Minimal example: app/lib/post-to-slack.ts This omits getTokenResponseReturns the full response: the token, its expiry, the connector identity, and provider metadata. Use when your app needs more than the raw string, for example to display the connected workspace name or to decide whether to surface a re-authorize button. Rate limit: 200 requests per minute per team. signature Minimal example: app/lib/inspect-token.ts getConnectorMetadataReturns the connector's stable public metadata and provider-specific configuration. Use it when your app needs connector setup values in addition to a runtime token, such as a Snowflake account identifier. Rate limit: 200 requests per minute per team. signature app/lib/snowflake.ts The calling project and environment must be linked to the connector, just as they must be for startAuthorizationStarts a user consent flow and returns the URL to send the user to. Call it when signature To send the user back to your app after the flow ends, pass app/lib/consent.ts All
MCP eventsUse Import the direct APIs and delivery verifier from Shared request parametersDirect API methods take
All direct methods accept
Non-success API responses use the Connect error types. experimental_discoverEventsDon't call this function. Vercel Connect does not support the operation, so the call always fails. To see which events a server offers, call experimental_listEventTypesLists event definitions from the authorized server's catalog.
Each descriptor includes experimental_subscribeCreates a subscription and returns
Pass Creation can return experimental_listSubscriptionsLists subscriptions for the connector, account, and calling project/environment. The list is not scoped to a conversation or the subscription's
Pass experimental_getSubscriptionRetrieves a subscription by ID after checking access.
Pass experimental_unsubscribeStops local delivery and attempts provider cleanup.
Pass experimental_unsubscribeAllDon't call this function. Vercel Connect does not support a bulk stop, so the call always fails. To stop every subscription for a connector and account in the calling project and environment, call Experimental_McpSubscriptionDirect subscribe, get, and unsubscribe calls return this object. Direct list calls return an array of these objects.
See subscription lifecycle for status handling and cleanup behavior. experimental_connectEventsAdapterReturns an
The returned adapter implements synchronous The bound operations are:
The adapter uses the same subscription lifetime rules as the direct API. All operations return promises. Event catalog discovery remains an MCP request through the AI SDK client. Subscription operations use Connect. See the AI SDK subscription example. experimental_verifyMcpEventAuthenticates a forwarded Connect delivery and validates its envelope: signature
The function consumes the request body. It does not acknowledge delivery, deduplicate it, or run application work. Durably accept by
Verification failures throw
Reject invalid deliveries. Treat configuration and verification availability failures as retryable receiver failures rather than acknowledging or processing the event. See the receiver example. TypesConnectTokenParams
ConnectTokenSubjectA union with four variants: @vercel/connect
The SDK infers ConnectTokenResponse
ConnectorMetadata
ConnectOptions
revokeTokenRevokes the provider grant for a connector subject and clears the SDK's in-process token cache. Provider support for revocation varies; see Revocation. Rate limit: 50 requests per minute per team (write operation). app/lib/disconnect.ts deleteTokenCacheEntryRemoves one cached token without revoking its provider grant. Pass the same connector and complete request parameters that you used for app/lib/retry-token.ts CachingThe SDK maintains an in-process LRU cache with a maximum of 100 entries, keyed by the connector and the full request params. Cached tokens are reused on subsequent calls until they fall inside the Pass ErrorsThe SDK throws typed error classes you can match on with
CLI parityMost SDK calls have a Next steps
|
|
On this page DEPLOYMENT_NOT_FOUNDThe 404 DEPLOYMENT_NOT_FOUND: Not Found TroubleshootAgent Prompt I'm encountering an error and reviewing the docs at https://vercel.com/docs/errors/deployment\_not\_found.md to understand what's happening. Please help me resolve this by: 1. **Suggest the fix**: Analyze my codebase context and propose what needs to be changed to resolve this error. If you do not have access to my codebase, ask me for the codebase and try to fix the error based on the information you have. 2. **Explain the root cause**: Break down why this error occurred: - What was the code actually doing vs. what it needed to do? - What conditions triggered this specific error? - What misconception or oversight led to this? 3. **Teach the concept**: Help me understand the underlying principle: - Why does this error exist and what is it protecting me from? - What's the correct mental model for this concept? - How does this fit into the broader framework/language design? 4. **Show warning signs**: Help me recognize this pattern in the future: - What should I look out for that might cause this again? - Are there similar mistakes I might make in related scenarios? - What code smells or patterns indicate this issue? 5. **Discuss alternatives**: Explain if there are different valid approaches and their trade-offs My goal is to fix the immediate issue while building lasting understanding so I can avoid and resolve similar errors independently in the future. Show less To troubleshoot this error, follow these steps:
|
Signed-off-by: zskbot <241179063+zskbot@users.noreply.github.com>
Merge pull request #18 from OCodejs/fix/base-branch-workflow-errors
| on: | ||
| workflow_dispatch: |
There was a problem hiding this comment.
Removing inputs.namespace leaves the Endor scan reading an input that maintainers can no longer supply. It also breaks the existing assertion in tests/repository-config.test.mjs:27, so the main CI job fails during npm test. Restore the required namespace input.
| on: | |
| workflow_dispatch: | |
| on: | |
| workflow_dispatch: | |
| inputs: | |
| namespace: | |
| description: Endor Labs tenant namespace | |
| required: true | |
| type: string |
Prompt To Fix With AI
This is a comment left during a code review.
Path: .github/workflows/endorlabs.yml
Line: 3-4
Comment:
**Missing namespace breaks CI**
Removing `inputs.namespace` leaves the Endor scan reading an input that maintainers can no longer supply. It also breaks the existing assertion in `tests/repository-config.test.mjs:27`, so the main CI job fails during `npm test`. Restore the required namespace input.
```suggestion
on:
workflow_dispatch:
inputs:
namespace:
description: Endor Labs tenant namespace
required: true
type: string
```
---
For each issue above, determine whether it is valid and should be fixed. If so, fix it directly.| actions: read | ||
| id-token: write | ||
|
|
||
| jobs: |
There was a problem hiding this comment.
Removing the permissions block also removes id-token: write, which the Endor action needs for its default GitHub OIDC sign-in. This workflow supplies no alternative credentials. Manual scans therefore cannot sign in even after the namespace input is restored. Keep the explicit permissions, including security-events: write for the report upload.
| jobs: | |
| permissions: | |
| contents: read | |
| security-events: write | |
| actions: read | |
| id-token: write | |
| jobs: |
Prompt To Fix With AI
This is a comment left during a code review.
Path: .github/workflows/endorlabs.yml
Line: 6
Comment:
**Endor scans cannot sign in**
Removing the permissions block also removes `id-token: write`, which the Endor action needs for its default GitHub OIDC sign-in. This workflow supplies no alternative credentials. Manual scans therefore cannot sign in even after the namespace input is restored. Keep the explicit permissions, including `security-events: write` for the report upload.
```suggestion
permissions:
contents: read
security-events: write
actions: read
id-token: write
jobs:
```
---
For each issue above, determine whether it is valid and should be fixed. If so, fix it directly.
Purpose
Prerequisite repair for PR #17. GitHub's
pull_request_targetevent executes workflow definitions from the base branch (main), so fixes in PR #17 alone cannot resolve these two failing checks.Verified root causes
redhat-actions/crda(repository not found)..github/labeler.ymldoes not exist onmain, so actions/labeler returns 404.Changes
No merge or Production deployment has been performed. After this prerequisite is merged, PR #17 can rerun its base-branch checks.
Restore the Endor namespace input and required permissions before merging.
Fix with agent prompt
Summary
This diff makes selected workflows manual-only and removes Endor Labs inputs and permissions.
Reviews (5) · Last reviewed commit: "Merge branch 'main' into fix/base-branch..." · Reviewed by Greptile