83,000 Laravel packages on Packagist. Only 19,000 are healthy or medium. The rest are old, stale, or abandoned.
AI can find you a package in seconds. What it cannot do is tell you whether it is still alive.
#Laravel #PHP #OpenSource
Remote
Daniel
@danielpetrica@infosec.exchange
mastodon 4.8.0-alpha.3+glitchAndrei-Daniel Petrica.
Web developer with 8 years of experience, trying to #BuildinPublic @LaraPlugins@phpc.social
Topics:
#Php, #Laravel, Javascript, Sql, #Docker, #Traefik, Japan, Anime
#fedi22
212 Followers
558 Following
44 Posts
Joined October 23, 2023
Website:
Building:
Laraplugins.io:
GitHub:
Gravatar:
Wrote a detailed performance post about the LaraPlugins infrastructure audit I did this week.
An 8GB VPS at 90% memory. 12 Horizon supervisors with unlimited jobs. 50,000 Cloudflare purge jobs in one cycle.
The 12 fixes that brought it back under control, with real numbers:
https://danielpetrica.com/laraplugins-performance-audit/
If you run a Laravel app with Horizon, some of these might hit close to home.
The most embarrassing fix in my performance audit this week was a single-column index I forgot to add. The most impactful was a JSON column exclusion on a read-heavy table.
12 fixes total. Every Laravel dev running a busy app will find something relevant.
Installing a New Relic agent takes 5 minutes. Configuring, calibrating, and testing the 5 alerts your SaaS actually needs takes 3 to 5 hours of focused work.
Most teams never get past the install step. They have data flowing but no real alerting. If your team is too busy shipping features, I handle the full setup in one week: danielpetrica.com/newrelic/
An expired certificate is never a complex failure. It's just a calendar problem nobody automated. Code is fine, servers are healthy, but every visitor gets a security warning.
Run a daily Synthetics Certificate Check in New Relic. Alert 14 days out if you use auto-renewals, 30 days out for manual setups. Takes 5 minutes, eliminates the dumbest kind of outage.
Free guide: danielpetrica.com/newrelic/
Server bills for LaraPlugins started creeping up. I did not want to put ads on the site or sell user data. So I added a curated tools section with things I personally used building the platform.
Every result is transparent about whether it is an affiliate link. Plugin health scores stay objective. The recommendations are my honest picks.
If you run a community project, how do you handle running costs?
A very persistent LLM scraper just gave my website a free stress test: 10k+ requests to a non-existent endpoint in under 24 hours
The result? Zero downtime. Zero performance degradation. And if i did not check the analytics I wouldhave missed it.
This is why I advocate for modern stacks like Traefik paired with FrankenPHP. The ability to handle high concurrency with minimal resource usage is a game changer for #PHP applications.
#developer #programming #technology #business
One lesson from building the proxy: Storage::download() and streamDownload() buffer the entire file into memory. That gets expensive with large images. Switching to readStream() + response()->stream() sends the file directly from S3 to the client, no buffering. Small detail, big impact when you are serving media at scale.
It is a statistical tie for the top frustration in the #Laravel ecosystem.
70% of developers cited lack of docs or usage examples as their biggest pain point. It ranks exactly as high as abandoned packages.
This sends a clear message to authors: You can write the cleanest, most performant #PHP code in the world, but if your Readme doesn't explain "how to," the community will view it as broken.
I dive into the full developer wishlist in my report: https://coz.jp/knr8Tr
#software #developer
Over the last 7 days LaraPlugins Had more than 18k requests made to its MCP Server.
Let's help your app too with extended health information data
https://laraplugins.io/mcp
A new chapter of my life has begun.
Hi everyone, since September I’ve decided to dedicate myself exclusively to freelance work. I know it will be a challenge, but I feel it’s the natural evolution of my career.
For the first time, I’m taking on clients for web development (Laravel, PHP, TypeScript, and Vue) as well as DevOps consulting (Docker, CI/CD, and Traefik).
I’ll also continue to offer freelance consulting on New Relic monitoring, something I’ve already done in the past.
If you have interesting projects or need targeted consulting, you can reach out to me here or through my website danielpetrica.com
Thank you all for your attention,
Have a great weekend
#freelance #webdev #developer #laravel #newrelic #vue_js
HI everyone!
I am a Senior Full Stack Developer and DevOps consultant with 8 years of experience and I currently have full time availability from February (next monday) onward.
I help teams design, build, and monitor scalable Laravel and Vue.js applications, from MVPs to long term production systems.
I work hands on with Laravel, Vue.js, FilamentPHP, Traefik, Docker and FrankenPHP, with a strong focus on clean code and observability (New Relic).
Some example projects I have worked on:
Custom platform for an Italian utility company.
Payment request and information portals for Italian municipalities. (pagoPa)
Custom SaaS applications built on Laravel and Vue.js
Monitoring and observability setups for consumer and security companies using New Relic
Monitoring and code quality improvements for a large Italian consulting enterprise
If you are looking for someone who can both ship features and keep your systems observable, secure, and reliable, feel free to reach out and share what you are working on.
#programming #fedihire #freelance #business #developer #laravel #php #vuejs #typescript #docker #mysql
I saw a direct debate in the survey responses regarding the frontend stack.
When asked what they would improve about the ecosystem, requests for "More #Livewire support" sat right next to simple demands to "Get rid of Livewire."
The community has strong, mixed feelings about the TALL stack versus other approaches.
Beyond the framework wars, the survey revealed deep insights into what #Laravel devs actually want from the ecosystem.
Read the community responses here https://coz.jp/knr8Tr
In the incoming wave of #AI dev, having a trustworthy directory is critical.
For agents to build reliable software, they need to be able to pick healthy packages.
If the data source is cluttered with abandon-ware, AI will produce broken implementations.
This is a motivation behind my work on https://LaraPlugins.io. We aren't just organizing packages for humans; we are organizing them for the next automated tools too.
My mcp server to search #Laravel packages is finally working correctly could you give it a try?
Find the instructions for it here: https://laraplugins.io/mcp
#programming #php #developer
The Italian term "grilletto facile" (quick trigger) perfectly describes how a modern #developer selects dependencies.
I asked 200 #Laravel devs how long they spend researching before typing "composer require".
61% spend between 5 and 30 minutes.
19% spend less than 5 minutes.
For the majority, the selection criteria must be scannable instantly: update recency, docs quality, and popularity. We rarely do deep code audits anymore.
See the full breakdown: https://coz.jp/knr8Tr
We have all experienced searching for functionality and finding 5 packages that haven't been updated since #Laravel 8
In my recent survey of the ecosystem, 70% of devs cited "abandoned or outdated packages" as their number 1 frustration. It is a statistical tie with poor docs.
The vibrancy of our community creates noise.
Filtering through that noise to find maintained tools is becoming a significant operational cost.
The full data here https://coz.jp/knr8Tr
decided to open an account for the #saas I'm trying to #buildinpublic find it here if you want to suport: @coz_jp@mastodon.social
My objectives for the weekend:
Finish API support for @coz_jp@mastodon.social
Study Japanese conversations
Finish an article on umami and Laravel.
Study MCP and Laravel again
Scaling a Laravel app brings fun challenges. Today, Laraplugins crossed 50,000 plugins!
This milestone highlighted some inefficiencies, causing FrankenPHP to exhaust its default 256MB memory limit and throw 500 errors.
Initially tried mounting a php .ini via docker compose volumes, but ultimately resolved it by properly copying the config directly in the Dockerfile.
Apologies for the 1-hour intermittent downtime while debugging, but we are back up and stronger than before
Started a blog for LaraPlugins. First post: our 151.37 HackerNoon Proof of Usefulness score breakdown.
I'll be writing openly about building a community tool. Architecture decisions, traffic numbers, things that break, things that work. One developer avoiding an abandoned package because of something here is enough.
Coming: MCP deep dive, what broke, full survey results.
https://laraplugins.io/blog/laraplugins-proof-of-usefulness-hackernoon-2026
#Laravel #PHP #OpenSource #buildinpublic
One thing I learned the hard way after nuking my Mastodon instance: automated offsite backups are non-negotiable.
I updated my Docker Compose backup guide with the renamed nfrastack image, S3 streaming strategies, and a new configurator tool. You describe your databases, it writes the YAML and env file. Handy when you need to protect multiple projects at once.
https://danielpetrica.com/easy-database-backups-with-docker-compose/
"If you could wave a magic wand and improve one thing about the #Laravel ecosystem today, what would it be?"
I asked this to 200 #developer. The recurring theme was a desire for a centralized, standardized way to evaluate packages.
"Standardize package info ..."
"Reduce the number of duplicate packages."
This feedback was the inspiration for https://LaraPlugins.io. We need a directory that focuses on health, not just download counts.
Full report https://coz.jp/knr8Tr
Quick reminder, remember to clear the image cache of docker if you do often deploys.
Server costs for LaraPlugins are adding up. Domains, hosting, CDN. So I added a recommended tools section with things I actually used during development.
Every recommendation is clearly tagged when affiliate. Plugin health scores stay completely separate.
Keeping a community project sustainable without compromising trust.
One hour of downtime costs more than a professional monitoring setup. Lost signups, churned trials, support overhead, team scrambling through logs while the outage compounds.
Full New Relic APM, database, logs, browser, and alerting configured and documented in one week: danielpetrica.com/newrelic/
#NewRelic #DevOps #SaaS
How much "baggage" is currently in your composer.json file?
I surveyed 200 active developers to gauge the average weight of a modern Laravel application. The results highlighted a distinct split in the community regarding the "Build vs. Buy" mentality.
Over half of the respondents (54%) fall into the 6–15 package range. Given standard best practices today involving tools like PHPStan, Pint, and Debugbar, this feels like the industry standard.
#Laravel #PHP #developer #programming #technology
When needing standard functionality like SEO or Media handling, what is your default reaction? Do you reinvent the wheel or look for existing solutions?
My recent survey shows a massive consensus: 96% of #Laravel devs immediately look for existing packages . Only a tiny fraction start by building from scratch.
This confirms that the modern dev is an integrator first. We focus on business logic and offload the plumbing to the ecosystem.
Read the analysis: https://coz.jp/knr8Tr
#programming
Laravel lets you monitor your configuration. But your dependencies?
You check a package once, install it, then never look again. Packages go stale quietly.
#Laravel #PHP #OpenSource
@LaraPlugins@phpc.social reached such a scale i want an automatic way to monitor its seo health so i cooked a new site.
looking to test it this week to test technical on page seo
Finally managed to export my 2fa tokens hold captive on Authy and move them to an opensource tool. I was missing having the commodity of the app on the desktop.
Alerting on "error rate > 5%" sounds sensible. But 2% errors at 3 AM is noise. 2% at peak traffic is a system failure. Use anomaly detection. Let New Relic baseline your data for 48 hours. Moderate sensitivity avoids phantom alerts. Guide: danielpetrica.com/newrelic/ #NewRelic #DevOps #monitoring
Replying to
@daniel@danielpetrica.com the plugin is not ready for use yet, but here is its page
https://laraplugins.io/plugins/danielpetrica/activitypub-for-laravel
That 4 downloads are impressive 🤣
Where do #Laravel devs actually find their tools? I was curious if my personal habit of just Googling everything was normal.
It turns out, it is.
74% of developers rely on Web Search
51% use Packagist directly
49% trust Newsletters/Curated News
Interestingly, GitHub Search (42%) and Socials (19%) ranked significantly lower. We seem to trust curated editorial content and algorithms much more than raw repository searches or social feeds.
See the full data:
https://coz.jp/knr8Tr
#developer
Replying to
@idiran@livellosegreto.it ciao prova a entrare nei gruppi di Modena e Reggio Emilia Full-Stack. Abbiamo molti sviluppatori locali e magari qualcuno li potrebbe indirizzarti nella direzione giusta. scusa il doppio post avevo usato account sbagliato
DB pool exhaustion: connections climb for hours, everything looks fine, then max_connections hits and your app collapses. Check yours: SHOW VARIABLES LIKE 'max_connections'. Default 151? Warning at 70% (~105), critical at 85% (~128). Track threadsConnected in New Relic. Setup guide: danielpetrica.com/newrelic/ #NewRelic #DevOps #MySQL
Gonna attend a php conference in Tokyo today without speaking enough Japanese.
Wish me luck please.
The packagist search Api only returns up to 100 pages at 100 elements per page.
even if there are more results in the set, per their count.
Updated my database backup guide for the first time since 2023. The tiredofit/db-backup image moved to nfrastack/container-db-backup, so the article now reflects the new name and the multi-job env variable format.
I also built a small configurator tool that generates docker-compose and .env files on the fly. Point it at your databases, pick your settings, copy the output. Saves a lot of repetitive editing.
https://danielpetrica.com/easy-database-backups-with-docker-compose/
#Laravel #DevOps #Docker
I'd like to have an orchestrator agent in my projects, letting him focus on controlling sub-agents and giving them directions.