In December I made a lot of optimization to increase throughput and system efficiency by batching up sql and redis commands, SQL index according to query and using a rate limiter internally.
Changelog for December 2025
Changed
- Storing and processing jobs in batches for better throughput
- Reduced DB calls in job submit worker by storing parent data in memory
- Using Zeroize to zero out the sensitive data in memory on drop
Changed
- Storing and processing jobs in batches for better throughput
- Reduced DB calls in job submit worker by storing parent data in memory
- Using Zeroize to zero out the sensitive data in memory on drop
January 1, 2026 at 12:18 PM
In December I made a lot of optimization to increase throughput and system efficiency by batching up sql and redis commands, SQL index according to query and using a rate limiter internally.
AgentFS by looks promising for local agent development
github.com/tursodatabas...
github.com/tursodatabas...
GitHub - tursodatabase/agentfs: The filesystem for agents.
The filesystem for agents. Contribute to tursodatabase/agentfs development by creating an account on GitHub.
github.com
December 16, 2025 at 1:15 PM
AgentFS by looks promising for local agent development
github.com/tursodatabas...
github.com/tursodatabas...
November 27, 2025 at 1:42 PM
Docker has put pull limits for unauthenticated and authenticated (free user) to 100 and 200 requests per 6 hour. One of the ways to not hit this without upgrading to pro account is to use digest has in the docker image name instead of tag.
November 21, 2025 at 4:30 AM
Docker has put pull limits for unauthenticated and authenticated (free user) to 100 and 200 requests per 6 hour. One of the ways to not hit this without upgrading to pro account is to use digest has in the docker image name instead of tag.
Data engineers are really loving DuckDB because it's like sqlite but very fast and flexible...
October 17, 2025 at 4:30 AM
Data engineers are really loving DuckDB because it's like sqlite but very fast and flexible...
If you use Linode or DigitalOcean, Daestro is a great fit. It has first-class support for their APIs, allowing you to orchestrate workloads on these providers just as easily as on AWS. It’s great to see support for alternative clouds.
October 16, 2025 at 4:30 AM
If you use Linode or DigitalOcean, Daestro is a great fit. It has first-class support for their APIs, allowing you to orchestrate workloads on these providers just as easily as on AWS. It’s great to see support for alternative clouds.
Ever wished you could prioritize your background jobs? Daestro's Job Queues let you do just that. Set a priority from 1-1000 (lower is higher!) to ensure critical tasks are processed first. It's simple but powerful control.
October 15, 2025 at 4:30 AM
Ever wished you could prioritize your background jobs? Daestro's Job Queues let you do just that. Set a priority from 1-1000 (lower is higher!) to ensure critical tasks are processed first. It's simple but powerful control.
Want faster PostgreSQL ops? Check out Pipelining! 🚀 It's a game-changer for throughput because your client, network, and server run tasks *at the same time*. I've used it for bulk inserts and the packet reduction is wild!
October 14, 2025 at 4:30 AM
Want faster PostgreSQL ops? Check out Pipelining! 🚀 It's a game-changer for throughput because your client, network, and server run tasks *at the same time*. I've used it for bulk inserts and the packet reduction is wild!
The way Daestro combines cloud and self-hosted resources is seamless. You can have a single Job Queue that contains Compute Environments from both AWS and your own on-prem servers. The platform handles job distribution automatically. True hybrid cloud!
October 13, 2025 at 4:30 AM
The way Daestro combines cloud and self-hosted resources is seamless. You can have a single Job Queue that contains Compute Environments from both AWS and your own on-prem servers. The platform handles job distribution automatically. True hybrid cloud!
A fun fact about Daestro: you can run jobs on your own machine! 💻 The platform allows you to configure a "Self-hosted Compute" environment. Your laptop or a server in your office can become a worker node in your orchestration setup.
October 12, 2025 at 4:35 AM
A fun fact about Daestro: you can run jobs on your own machine! 💻 The platform allows you to configure a "Self-hosted Compute" environment. Your laptop or a server in your office can become a worker node in your orchestration setup.
Want to prevent runaway cloud costs? Daestro's Job Queues have a "Max Idle Time" setting. This automatically terminates a compute instance if it has been idle for a specified number of seconds. 💸 A smart, built-in feature for cost optimization.
October 11, 2025 at 3:39 AM
Want to prevent runaway cloud costs? Daestro's Job Queues have a "Max Idle Time" setting. This automatically terminates a compute instance if it has been idle for a specified number of seconds. 💸 A smart, built-in feature for cost optimization.
While Docker is king, sometimes a simple bash script is all you need. Daestro lets you run arbitrary bash scripts as jobs within a default `ubuntu:24.04` container. 📜 It's incredibly handy for those quick, one-off automation tasks. Do you still rely on bash? #Bash #Scripting
October 10, 2025 at 3:39 PM
While Docker is king, sometimes a simple bash script is all you need. Daestro lets you run arbitrary bash scripts as jobs within a default `ubuntu:24.04` container. 📜 It's incredibly handy for those quick, one-off automation tasks. Do you still rely on bash? #Bash #Scripting
Daestro has a transparent cost model. You add your cloud provider's API key, and it provisions resources directly in your account. You pay your provider for usage, and Daestro only charges a platform fee. No hidden markups. Clear and simple. 👍 #FinOps #CloudCosts #AWS #Vultr
October 10, 2025 at 9:33 AM
Daestro has a transparent cost model. You add your cloud provider's API key, and it provisions resources directly in your account. You pay your provider for usage, and Daestro only charges a platform fee. No hidden markups. Clear and simple. 👍 #FinOps #CloudCosts #AWS #Vultr
Looking for a way to run compute workloads without getting tied to a single cloud? ☁️ Daestro is a cloud-agnostic platform letting you orchestrate jobs across AWS, Vultr, Linode, DigitalOcean, and even your own on-prem hardware. True flexibility. What's your multi-cloud strategy?
October 10, 2025 at 3:39 AM
Looking for a way to run compute workloads without getting tied to a single cloud? ☁️ Daestro is a cloud-agnostic platform letting you orchestrate jobs across AWS, Vultr, Linode, DigitalOcean, and even your own on-prem hardware. True flexibility. What's your multi-cloud strategy?
Workflow platform n8n just announced a massive $180M funding round! 🤯 This isn't just about money; it highlights the soaring demand for powerful, flexible automation tools, especially those embracing open core principles.
October 9, 2025 at 3:39 PM
Workflow platform n8n just announced a massive $180M funding round! 🤯 This isn't just about money; it highlights the soaring demand for powerful, flexible automation tools, especially those embracing open core principles.
Database Tip Alert! 🚨 When a user leaves a text field empty, your server sees an empty string (""), not a NULL value. I spent hours debugging this early in my career! 🤦 Fun fact: Always check for "" and convert it to NULL before saving. 90% of data cleanup issues start here!
October 9, 2025 at 7:31 AM
Database Tip Alert! 🚨 When a user leaves a text field empty, your server sees an empty string (""), not a NULL value. I spent hours debugging this early in my career! 🤦 Fun fact: Always check for "" and convert it to NULL before saving. 90% of data cleanup issues start here!
Now you can run multiple Jobs on single server with Daestro. Read more here: daestro.com/blog/run-mul...
Run multiple Jobs on single Server with Daestro - Daestro
Steps to run multiple jobs on one compute spawn with CPU and Memory quota
daestro.com
September 1, 2025 at 9:41 AM
Now you can run multiple Jobs on single server with Daestro. Read more here: daestro.com/blog/run-mul...
Nice write up, really helpful
🏗️ Organizing Larger Projects #rustlang
How do large projects like bevy or the rust compiler get broken down into manageable chunks? How do they keep the codebase modular and maintainable, and easy to work on? Do you have a single main file with 500+ LOC?
Let's learn how to use workspaces🧵👇
How do large projects like bevy or the rust compiler get broken down into manageable chunks? How do they keep the codebase modular and maintainable, and easy to work on? Do you have a single main file with 500+ LOC?
Let's learn how to use workspaces🧵👇
July 28, 2025 at 5:51 PM
Nice write up, really helpful
Vultr is an exceptional cloud service provider, offering highly cost-effective VPS solutions. While they don't have built-in support for batch processing, you can utilize DaemonStack to execute batch tasks on Vultr. Find out more here:
How to run batch jobs on Vultr - DaemonStack
This guide will show you how to run batch jobs on Vultr
daemonstack.com
February 20, 2025 at 1:07 PM
Vultr is an exceptional cloud service provider, offering highly cost-effective VPS solutions. While they don't have built-in support for batch processing, you can utilize DaemonStack to execute batch tasks on Vultr. Find out more here:
Introducing DaemonStack, a cloud agnostic platform that helps you run compute workload across cloud provider.
Go to: daemonstack.com
Go to: daemonstack.com
February 19, 2025 at 11:43 AM
Introducing DaemonStack, a cloud agnostic platform that helps you run compute workload across cloud provider.
Go to: daemonstack.com
Go to: daemonstack.com
Reposted by Vivek Shukla
I miss games that are about 8-20h long and have linear gameplay/stories. Not every game has to be open world and 60+h long😭
I'm so sick of it. If you don't have extensive worldbuilding to fill that world just tell your story in a (semi) linear way 🥲
I'm so sick of it. If you don't have extensive worldbuilding to fill that world just tell your story in a (semi) linear way 🥲
February 17, 2025 at 10:00 AM
I miss games that are about 8-20h long and have linear gameplay/stories. Not every game has to be open world and 60+h long😭
I'm so sick of it. If you don't have extensive worldbuilding to fill that world just tell your story in a (semi) linear way 🥲
I'm so sick of it. If you don't have extensive worldbuilding to fill that world just tell your story in a (semi) linear way 🥲
Added support for Vultr on @daemonstack.bsky.social, now you can run your compute workload on Vultr servers.
February 8, 2025 at 3:43 PM
Added support for Vultr on @daemonstack.bsky.social, now you can run your compute workload on Vultr servers.
If you are using SPA web app and calling external api, you probably have seen that it takes longer than usual to get the response back. This is due to CORS. Browser does preflight check before each request to make sure the method and request headers are allowed.
February 8, 2025 at 2:34 PM
If you are using SPA web app and calling external api, you probably have seen that it takes longer than usual to get the response back. This is due to CORS. Browser does preflight check before each request to make sure the method and request headers are allowed.