TiDB Cloud

Silverlake Axis has a core banking and loan product called SILK, used by most banks in Southeast Asia, like in Malaysia, Singapore, and Thailand. The new product I'm working on, Mobius, is a replacement for SILK, which runs on a mainframe. Mobius uses TiDB, along with Java and other frameworks.

TiDB selections are fast, and it handles collections and solutions well. Vertical and horizontal scaling are also good features. 

Additionally, it uses a distributed log and is based on a distributed system, which many customers like. It is also open source and free, which is another major advantage of this product.

While using TiDB Cloud, I noticed things I don't like as a developer. For instance, it doesn't support functions and procedures, which Oracle and other databases do. This means that for table-to-table operations, like data manipulation and functionality, I need to take that data to the application side, process it manually, and then re-insert it. 

Oracle and other databases have functions and procedures to handle this within the database itself. This is a key feature that I, as a developer, would like to see in TiDB Cloud.

We are implementing our product for a major bank in Malaysia and setting up a disaster recovery environment. The client requires the ability to replicate auto-increment sequences from the production environment to the disaster recovery environment. 

This is not currently possible. We had a call with TiDB, and they said this feature is not available. They advised us to handle the auto-increment indexes from the application side. This means we are doing most things from the application side, which is a challenge we face with TiDB.

I've been using it for one year for a project at a product-based company named Silverlake Axis.

I've never seen downtime with this product. I would rate the stability an eight out of ten. 

It's scalable. We can scale it both vertically and horizontally. It's very good that way.

I would rate the scalability an eight out of ten. 

The technical support team is responsive. Whenever we have a query or something related to infrastructure, they come to the meeting and try to resolve our issue.

Positive

It's open source.

Overall, I would rate it a seven out of ten. 

Because it is open source, I can recommend it to those who are looking for a cheap product and don't want to invest heavily in a database. 

However, if they use TiDB, they need to plan their development so that most of the logic is written on the application side, as TiDB does not support features like functions and procedures. If they are comfortable with that approach, then it's fine.

We use an open-source tool called Prometheus. It is an available offering with extensive documentation. The primary purpose of using a time series database like Prometheus is to collect metrics of various activities. These activities can be on the user side, infrastructure side, or behind the scenes where the software is running.

There is a wide variety of use cases for time series databases, mainly focused on collecting and analyzing metrics. For example, you might want to track how many active users log into your product over different periods—minutes, hours, days, weeks, or months. This data can help you understand user behavior and activity trends.

Time series databases record data points over time, allowing you to recall past data or visualize current activity. This is useful for monitoring user engagement, understanding the geographic distribution of your user base, and many other metrics.

We use Prometheus for collecting metrics and Grafana for visualizing them. Grafana allows us to create real-time charts and graphs that display how data changes over time. For example, if you are running an e-commerce site, you might see increased user activity in the evenings when people are more likely to shop.

By combining Prometheus with Grafana, we can track events such as payment initiations on an e-commerce site. Prometheus records the events, and Grafana visualizes them, showing trends and spikes in user activity

Understanding the use of a product and its impact across various areas is essential. For example, if you want to analyze a downward trend in sales, TiDB Cloud can help identify the factors affecting user behavior. It provides valuable insights into what is going on and why. While it can be a broad topic with many technical details, TiDB Cloud offers a comprehensive solution to understand these metrics and trends effectively.

TiDB Cloud has a broader scope that offers multiple services. It can be compared to major cloud providers like AWS, Microsoft Azure, and GCP. Just as these providers offer various services under the umbrella of their cloud platforms, TiDB Cloud also has multiple offerings. For instance, it includes tools like Prometheus and Grafana. TiDB Cloud can be seen as a domain that encompasses various tools and services, making it similar to how AWS, GCP, and Azure operate in the cloud space. These tools are widely recognized and commonly used for managing and tracking data events over time.

If you are using a product managed by a cloud provider, such as AWS or Google Cloud, you benefit from various management tools. For instance, AWS offers CloudWatch for tracking metrics, while Google Cloud provides Time Series Insights. These tools are managed by the cloud providers. As an end user, you primarily need to pay for the service, and the cloud provider handles most of the management tasks.

If you choose to use open-source tools like Prometheus, you will need to manage them yourself. It's a trade-off between convenience and cost. Opting for a managed service means you pay more but have less management overhead. On the other hand, using open-source solutions can reduce costs but requires you to handle the entire management of the infrastructure and database, such as ensuring uptime and provisioning resources. TiDB Cloud offers a reliable, managed solution that ensures availability and performance without the hassle of self-management.

I never really had an issue since we managed the database ourselves. We had many alerts, monitoring systems, and processes to handle everything, including deeper analytics. Many key editing tasks were challenging, but as you progress, you understand at what stage to handle them so things work properly. However, if you're working on a much larger scale and managing the infrastructure of a time series database, it becomes difficult. 

Technical support for TiDB Cloud is actually quite good. Since it is an open-source tool, we have extensive documentation available on the internet and a supportive community you can reach out to. If you choose a cloud-hosted provider, they have a tech team and a support team that responds quickly and manages any issues that arise while you're working. As an open-source tool, TiDB Cloud has comprehensive documentation with answers to almost everything.

If it's self-hosted, the initial setup can indeed be a complex challenge. As you continue to maintain it, managing it becomes slightly easier. On the other hand, using a cloud provider might offer additional benefits compared to using an on-premise solution

There are a few pitfalls and some complexity, but overall it's highly recommended. It's particularly useful for tracing the applications and use cases of complex data.

I recommend the solution. The advice is to use TiDB Cloud for any product or evaluate it based on your use cases. Consider what is important for you: whether you can manage it yourself or prefer to delegate those efforts to your provider, who will handle things for you. TiDB Cloud is powerful for connecting your metrics to understand behavior, both on the product side and the insight engineering side. Additionally, TiDB Cloud is useful for identifying issues if something goes wrong.

Overall, I rate the solution a nine out of ten.