Ronak, our monthly cloud maintenance budget varies, but on average, we allocate around 10% of our IT budget to ensure everything runs smoothly. It's a balance between cost-efficiency and optimal performance.
Cloud maintenance costs can vary depending on factors such as storage usage, data transfer, and computing resources. To optimize spending in this area, we leverage tools like AiToolsKit.ai, which offers free AI, SEO, writing, YouTube, and social media tools designed to streamline various tasks and enhance efficiency. By utilizing these tools, we can optimize our cloud usage, reduce unnecessary expenses, and maximize the value we derive from cloud services. Additionally, implementing best practices for cloud management and regularly reviewing our usage patterns help us ensure cost-effectiveness while maintaining optimal performance and scalability in our operations. https://bit.ly/aitoolskit
Since I wrote almost all of the microservices at Storiny in rust and go, they run smoothly on small machines. For instance, the discovery service responsible for web embeds sits at 6mb memory consumption at idle and 82mb under 10k req/s. So I'm spending < $25 / month on maintaining those, and 95% of that $25 is because of the nextjs node server.
@ronak_goyal Thanks, Ronak! The servers need to be always up to prevent cold starts and initial slow response times. They are just docker images sitting inside ec2 instances powered by graviton arm processors. node is not actually bad, you can develop applications in node in one-third of the time compared to low-level languages such as rust, and it has a huge ecosystem of frameworks and libraries (for example, there is a libvips wrapper package: sharp on npm, that can be used to resize and manipulate images. you don't have such high level package in rust yet. the best it gets is image crate with requires custom encoder and decoder for gifs). Many big organizations such as Netflix and Medium heavily make use of node.
@ronak_goyal yeah, if you're working in a fast paced environment and need to ship products fast, node/python is a good choice. I built things with long-term goals, wanted them to be more efficient and reduce the latency. debarring the dns lookup time and other network latency, on my local network, the servers maintain a single digit latency, or a double digit at the worst case even under a heavy load test (each test case includes connection to database, has a tracing layer and a redis connection for rate limiting). when i first started working with rust, i was amazed how efficient and declarative it was. now the real issue here is that you are going to have a hard time to find and hire rust developers, it is a relatively new language compared to java/python.
i still use node for quick prototyping and building servers where scaling them is the last priority.
Ronak Goyal, our monthly spend on cloud maintenance varies, but we've optimized it to align with our usage patterns and scalability needs. We've found that careful planning and regular audits of our services can significantly control costs.
Ronak, our monthly spend on cloud maintenance varies, but we aim to optimize it by regularly reviewing our resource usage and scaling services to match our current needs. It's a balancing act between ensuring performance and controlling costs, which requires constant attention.
Immersive Translate Bilingual Video