Skip to content

Dependabots alerts fixes - #5517

Merged
narcisonunez merged 6 commits into
canaryfrom
dependabots-alerts
Oct 14, 2026
Merged

narcisonunez merged 6 commits into
canaryfrom
dependabots-alerts

Conversation

@narcisonunez

@narcisonunez narcisonunez commented Sep 23, 2026 •

Copy link
Copy Markdown
Collaborator

What is this PR about?

This PR will fix almost all dependabots alerts. The remaining ones will be addressed later when we implement some new features from better-auth. The new version changes couple things

Checklist

Before submitting this PR, please make sure that:

Issues related (if applicable)

Screenshots (if applicable)

RetriggerConfidence Score: 4/5

The PR appears generally safe to merge, but the unscoped deepmerge-ts major-version override should be corrected or validated against Prisma first.

Summary

This PR updates direct and transitive JavaScript and Go dependencies to remediate Dependabot alerts, refreshes the lockfiles, and migrates the renamed Faker username API.

  • Upgrades runtime libraries across the API, dashboard, scheduler, server, and monitoring workspaces.
  • Adds root pnpm overrides for patched transitive dependency versions.
  • Updates template username generation for Faker v10.
  • One override crosses a dependency's declared major-version boundary and should be scoped or validated before merging.

Reviews (1) · Last reviewed commit: "fix(deps): resolve open Dependabot alert..."

@narcisonunez
narcisonunez marked this pull request as ready for review September 24, 2026 01:57
Comment thread package.json Resolved
narcisonunez added a commit to Dokploy/templates that referenced this pull request Sep 29, 2026
…parsing

Dokploy is moving to toml 4.2.0 (Dokploy/dokploy#5517) and fetches template.toml
from this repo at runtime, so these templates would fail to parse there.

- signoz, supabase: escape invalid backslash sequences (\{ \$ \( \s ...) as \\
  so the parsed value keeps the literal backslash exactly as toml 3 produced it
- timescaledb, valkey: drop `mounts = []` that was redefined by [[config.mounts]]

Parsed output is deep-equal to before under toml 3 for all 532 templates.
narcisonunez added a commit to Dokploy/templates that referenced this pull request Sep 29, 2026
Dependencies:
- app: react-router-dom ^7.18.0, vite ^6.4.3, yaml 2.8.3
- app (transitive, lockfile only): postcss 8.5.28, nanoid 3.3.19,
  picomatch 2.3.2/4.0.7, browserslist 4.29.2, rollup 4.63.5,
  @babel/core 7.29.7, @babel/runtime 7.29.7
- build-scripts: toml 4.2.0 (matches Dokploy/dokploy#5517), yaml 2.8.3

Templates rejected by toml 4's strict parsing (Dokploy fetches template.toml
from this repo at runtime, so they would fail there once #5517 ships):
- signoz, supabase: write invalid escape sequences (\{ \$ \( \s ...) as \\
  so the parsed value keeps the literal backslash exactly as toml 3 produced it
- timescaledb, valkey: drop `mounts = []` redefined by [[config.mounts]]

Parsed output is deep-equal to before under toml 3 for all 532 templates.
@agentHits

Copy link
Copy Markdown
Contributor

Thanks for this, @narcisonunez! It already clears most of the Dependabot alerts. I ran the npm advisory audit against this branch's lockfile (head 5e0b021) today: 220 advisories on canary, 30 here. Two of the remaining 30 are critical, and everything below can be fixed with a version bump:

Package In this branch Advisory Fixed in
next 16.3.4 GHSA-vcvr-r3jv-pc5j, critical 16.3.6 (latest 16.3.8)
proxy-addr 2.0.7, via express 5 (@modelcontextprotocol/sdk, inngest, express-rate-limit) GHSA-jqcg-44mw-7w3h, critical, published Oct 5 2.0.8
nodemailer 9.1.1 GHSA-v53p-9fqp-m79j, GHSA-prgh-xp8r-p3m5, high, published Sep 29–30 10.0.6 (latest 10.0.15)
axios 1.18.0, via swagger-ui-react 7 advisories, high 1.20.0
source-map-js 1.2.1, via postcss GHSA-68fv-2mgg-jv7q, high 1.2.2

Suggested changes:

  • apps/dokploy/package.json: "next": "16.3.8"
  • apps/dokploy/package.json and packages/server/package.json: "nodemailer": "10.0.15". Per its changelog, the only breaking change in 10.0.0 is requiring Node.js 20+, and Dokploy is on 24.
  • root package.json, next to the existing tar / protobufjs overrides:
    "axios": ">=1.20.0 <2.0.0",
    "proxy-addr": ">=2.0.8 <3.0.0",
    "source-map-js": ">=1.2.2 <2.0.0"

By my count that leaves about 10 advisories (3 high, 7 moderate) and no critical ones. The 3 high ones can't be fixed with a version bump yet. @better-auth/scim needs the better-auth 1.7 upgrade you mentioned, and braces / http-cache-semantics have no patched release upstream.

@Siumauricio, since you're the requested reviewer: is there a rough timeline for merging this? Days, weeks, or longer? It's been open since Sep 23 and fixes ~190 advisories, so it would help to know what to expect. Happy to help test the branch if that's useful.

@agentHits

Copy link
Copy Markdown
Contributor

You're doing a great job, and it's best not to rush into decisions—it'll quickly weed you out. But I hope someone knowledgeable in programming will step aside to develop and deploy code more quickly and improve this product, rather than dealing with stupid bureaucracy. Otherwise, it will be a month or more before you accept this PR.

@narcisonunez

Copy link
Copy Markdown
Collaborator Author

@agentHits Thank you for your support. We are focusing on community and open source features. We don't like to rush pushing changes without a proper validation from the internal team. We value all the effort people put into PRs and issues but we need to make sure we keep the product stable as much as we can.

That been said, we now have an internal policy to fix this dependabots alerts asap. At least the critical ones.

@agentHits

agentHits commented Oct 14, 2026 •

Copy link
Copy Markdown
Contributor

@narcisonunez @Siumauricio What kind of stability is this? If your project is riddled with bugs and security holes?

Step 1 - First, update the libraries to the latest versions and add support for all of these .
Step 2: reduce the panel's RAM consumption so it takes up less space and can run smoothly on low-end VPSes.
Step 3: focus on optimization and migrate the project to a modern BUN.

You haven't done anything meaningful to fix these three points since summer—your boss just gave up and that's it. You'd be better off focusing on stability and improving the product, fixing bugs, updating libraries, and rewriting the junk code to something more reliable and modern long ago. If you tell someone on the sidelines that a project collects $6,000 in donations, but is crawling along at a snail's pace—that's ridiculous, thanks to AI assistance—you could have done a lot.

Take my example - this is your fork https://github.com/agentHits/dokploy , and I'm making it my own. The fork has everything I described. If I were coding it myself, I would have done it even better and more. But at least I've eliminated security and the tricks of insecure libraries.

Anyway, it's worthwhile for me to split from you. I'm implementing what I need and monitoring security using AI, but I have more functionality and convenience than you.

@agentHits

agentHits commented Oct 14, 2026 •

Copy link
Copy Markdown
Contributor

@narcisonunez @Siumauricio It might even be more profitable for you to fork my project, if you agree. I can fully upgrade to your version—I'm 28 commits behind, and my fork is easier to fix because I've contributed a lot to the code, maybe 150,000 lines of code, maybe more.

But that's only if you want it. Until you do, your project will simply slow down. It's more profitable for me to update my fork than yours, because it's bureaucratic and a waste of time.

I'm surprised the owner can't hire staff who could run faster project updates and provide good support to resolve all issues in a timely manner—but frankly, I don't believe you'll improve, so I'll continue working on the fork https://github.com/agentHits/dokploy

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants