Important community update: Klaw's latest actions to enhance transparency​

We, as an integral part of the Klaw open-source project, are committed to upholding the true spirit of open-source values. To ensure transparency and foster trust within our community, we have an important update to share.

Discovery of legacy tracking code: We recently identified a piece of legacy, non-functional tracking code within our codebase. This code, originating from a previous version of our project, used Google tag (gtag) for analytics purposes. It was designed to track various user interactions on our site, including page views, scrolls, outbound link clicks, file downloads, video plays, site searches, and form interactions.

Scope and deactivation of the tracker: This tracking code was specifically configured to gather data from kafkawize.com. We want to reassure our community that this means that no data from any other host has been or is being collected, accepted or stored by this code. As of February 2022, this configuration has been completely removed.

Impact on users: We are pleased to inform you that the legacy code has not affected our newer Coral React UI. This interface, representing the latest development of our platform, remains free from any impact of this outdated tracking code.

In commitment to our community's transparency, we are releasing software updates 2.5.2 and 2.6.1. These versions aim to completely remove this legacy code, allowing our users to transition to a version where their privacy and trust are further safeguarded.

Your feedback is important. Stay updated and reach out to us via GitHub for any queries or suggestions.

This article explores the significance, evolution, and management of schemas in Klaw. It also covers how Klaw safeguards organizations against data loss by creating schema backups for disaster and failure recovery.

Schemas​

Data governance is becoming increasingly important for organizations that want to improve data quality, comply with regulations, and strengthen security. One way to achieve this goal is by setting up internal benchmarks, regulations for data, and schemas.

Schemas provide a clear understanding of what to expect from information, guiding their engagements and the relationships among objects. This guidance enables streamlined access, modification, and querying of data.

Furthermore, a schema outlines both the visual representation and structural organization of the data.

Apache Kafka® context​

Within the Apache Kafka ecosystem, schema registries such as Karapace and Confluent's Schema Registry play a key role in the management and validation of schemas. Furthermore, they provide a RESTful interface to manage schema and their evolution processes effectively.

Apache Kafka is a distributed streaming platform that allows producers to write data to topics and consumers to read from them. When there are structural changes to the data, it is important to establish an understanding between the producers and consumers to ensure that events are processed successfully.

The below image depicts the communication between Schema Registry, Apache Kafka, producers, and consumers.

Producers can register a schema based on a specific configuration, or they can use the REST API provided by the Schema Registry. During this process, events are serialized and validated against the currently registered schema. Subsequently, consumers deserialize the event and transform it into a standardized record in line with one of the registered schemas, as determined by the consumer's settings.

Schema id​

In Apache Kafka events, you can create schemas for both keys and values, allowing the generation of key schemas and value schemas. When a schema is registered, it is automatically stored in the default _schemas Apache Kafka topic and is assigned a unique identifier.

If the same schema is registered on another topic, it will link the previously registered schema using the same identifier. This identifier becomes an integer value generated upon successfully registering the schema.

To retrieve a schema based on an id: GET /schemas/ids/{int: id}/schema

Schema strategy​

Apache Kafka offers various strategies for schema registration, with the default being the TopicNameStrategy. In this approach, the subject name corresponds to the Apache Kafka topic name. In this strategy, typical subject names include "topic-key" or "topic-value."

Example of an Avro schema

{
  "type": "record",
  "namespace": "customer.account",
  "name": "CustomerAccount",
  "fields": [
    {
      "name": "Name",
      "type": "string"
    },
    {
      "name": "Age",
      "type": "int"
    }
  ]
}

Key schemas are usually strings by default, but you can register key schemas explicitly if needed. Apache Kafka supports alternative strategies, such as RecordNameStrategy and TopicRecordNameStrategy.

Schema evolution​

When you initially register a schema, the need for expansion and scalability arises as application requirements evolve. To manage this growth effectively, Apache Kafka consumers must be able to process events using both existing and newly registered schemas, thereby avoiding deserialization issues.

A key feature is accommodating multiple schema versions and allowing consumers to seamlessly deserialize Apache Kafka records according to the registered schema and event structure. This dynamic feature enables producers and consumers to generate and process events at scale while ensuring compatibility efficiently. It is important to note that each new schema version is assigned a unique identifier.

Supported compatibility types include BACKWARD, BACKWARD_TRANSITIVE, FORWARD, FORWARD_TRANSITIVE, FULL, FULL_TRANSITIVE, and NONE.

You can configure schema compatibility settings at both the subject and global levels. The system defaults to the global setting if you don't specify subject-level compatibility.

To check the compatibility status, you can use the following APIs:

• GET /config (global)
• GET /config/(string: subject)

Rest interface​

A REST interface is accessible for schema management, enabling tasks such as registering a new schema, submitting updates to it, deletion, and more.

• GET /subjects
• POST /subjects/(string: subject)/versions

Here is a detailed documentation :

• Karapace
• Confluent

Schema management in Klaw​

In Klaw, each Apache Kafka topic is associated with a specific owner, usually a designated team. The owner team is automatically granted the role of schema owner, giving them the authority to create new schemas and oversee their evolution.

Klaw ensures data consistency, reliability, and secure access, aligning with organizational policies. It also controls schema registration, permitting only the designated owner to manage schemas and preventing unauthorized modifications by other teams.

Request for a schema in Klaw​

After creating a topic, the designated team that owns the topic gains the exclusive right to register a schema for it. As the topic owner, you have the authority to approve, initiate, or reject any schema-related requests for that topic.

For more information on registering a schema see here

Watch the video below for a step-by-step demonstration of registering a schema in Klaw.

Schema migration from Schema Registry to Klaw​

Backing up schemas from the Schema Registry is essential for ensuring recovery in the event of server failures or unexpected disasters. While schemas are typically stored in a Apache Kafka topic with replication, there may be cases where retrieving them is difficult or impossible.

Klaw simplifies this by connecting directly to the Schema Registry server, retrieving all available schemas associated with topics, and integrating them into its metastore. This backup process is straightforward and takes only a few clicks, regardless of the volume of schemas or topics.

For more information on synchronizing schemas from the Schema Registry to Klaw, refer to sync-schemas-from-cluster

The following video demonstrates how schemas are synchronized to Klaw for a selection of topics already present in Karapace.

Schema migration from Klaw to Schema Registry​

After successfully backing up schemas in Klaw, you have the flexibility to register them either to a new Schema Registry cluster or integrate them into an existing one.

Select the schemas you want to sync with the cluster. Once you submit your choices, Klaw will register the selected schemas directly into your designated Schema Registry environment.

For more information on this synchronization process, refer to sync-schemas-to-cluster

The following video demonstrates migrating schemas to a Schema Registry, specifically focusing on a couple of subjects already stored in Klaw.

Conclusion​

Given the critical importance of data quality, protection, compliance, and management in every organization, Klaw provides a robust foundation for building a resilient data governance framework. With its automated features, Klaw empowers data teams to organize data, ensuring consistent accessibility, robust security, and unmatched reliability.

Background​

In today's dynamic and ever-changing digital landscape, maintaining the uninterrupted operation of applications is of the highest importance. This blog post explores the configuration of Klaw, a Java-based application designed specifically for Apache Kafka® governance.

Our primary focus is on exploring the configuration process for achieving High Availability (HA) in production environments. By implementing these measures, we can ensure continuous operation, enhance reliability, and deliver a seamless user experience, even during unexpected failures.

Why High Availability matters​

High Availability (HA) is a design principle to ensure operational continuity over a given period. For Klaw, being an essential tool in managing and monitoring Apache Kafka clusters, the application must be resilient to potential system failures. Achieving High Availability for Klaw means eliminating single points of failure and ensuring minimal or no downtime, thereby leading to a more reliable Apache Kafka governance. Unforeseen events and emergencies can cause server failures, affecting even the most robust systems. By supporting High Availability, Klaw is equipped to automatically recover from component failures, thereby reducing the impact of these events and providing users with a seamless experience.

Importance of High Availability in Klaw​

Apache Kafka clusters can be complex and substantial, making effective management and monitoring crucial. A failure in these operations can lead to significant issues. By configuring Klaw with High Availability, it becomes a reliable Apache Kafka governance tool that ensures uninterrupted operations, even in the face of underlying system failures. Deploying Klaw with High Availability minimizes or eliminates downtime and provides continuous support for Apache Kafka cluster management, enhancing overall system reliability. With this understanding of High Availability and its critical role in Klaw, let's explore the architectural design that supports it, the internal components of Klaw, and how to configure Klaw for High Availability.

What problem does High Availability for Klaw solve?​

Before diving into what High Availability is, let's understand why it's crucial for Klaw. As Klaw is essential in managing and monitoring Apache Kafka clusters, ensuring its resilience against potential failures is paramount. Downtime can have dire consequences ranging from slight inconveniences to lost revenue and a damaged reputation. High Availability for Klaw addresses these critical issues:

• Minimizing downtime: By eliminating single points of failure and ensuring redundancy in the system, HA for Klaw minimizes or eliminates downtime.
• Scalability: As the workload increases, Klaw can handle a higher number of requests, catering to a growing user base, thanks to the HA configuration.
• Data availability: Ensuring that the data is always available, even in the case of component failures, is crucial. HA ensures the data is replicated across different servers, safeguarding against data loss.
• Service continuity: In the event of a disaster or system failure, HA ensures that there is no interruption in service and the operations continue without a hitch.
• Enhanced user experience: Constant availability and reliability improve user experience, which is vital for customer satisfaction and retention.

What is High Availability?​

High Availability (HA) is a design approach that guarantees a certain degree of operational continuity during a given measurement period. It's about ensuring that applications remain available even if a critical component, such as a server, fails. Essentially, high availability eliminates any single point of failure and ensures minimal or no downtime. Why is High Availability important? Deploying applications with High Availability in production environments is essential for minimizing or eliminating downtime. Unforeseen events and emergencies can cause server failures, affecting even the most robust systems. HA systems are designed to automatically recover from component failures, reducing the impact of these events.

High Availability architecture​

Ensuring systems can handle higher workloads and substantial traffic is undeniably vital. However, it is equally essential to identify potential failure points and minimize downtime. A key component in achieving this goal is a highly available load balancer, which plays a significant role. It involves implementing a scalable infrastructure design that adapts to increased traffic demands. This often consists of employing a software architecture that surpasses hardware limitations.

Load balancing involves automatically distributing workloads across system resources, such as assigning various data requests to different servers.

In a highly available (HA) system, servers are organized in clusters and arranged in a tiered architecture to handle requests from load balancers efficiently. If a server within a cluster experiences a failure, a replicated server in a separate cluster can seamlessly take over the workload initially assigned to the failed server. This redundancy allows for failover, where a secondary component assumes the responsibilities of a primary component upon failure while minimizing any adverse effects on performance.

Understanding Klaw​

This section provides an in-depth overview of Klaw's architecture, the inner workings of its application, and the various perspectives it offers to users.

Klaw architecture​

Klaw is a web application developed using Java, compatible with JDK 17/19. The application's front end is initially built using AngularJS, but an upgrade to React JS is underway. The Backend development is carried out using Spring Boot. Klaw relies on an RDBMS-based data store for managing metadata. In its default configuration, Klaw employs a file based H2 database for storing metadata.

Klaw internal components​

Klaw is composed of two primary Java applications: the governance layer and the cluster layer.

• Governance layer

The governance layer is an integral part of Klaw, responsible for handling user interfaces and APIs.

User interface components: Allows users to interact with Klaw's features through a web interface. APIs and Authorization: The APIs in the governance layer are responsible for authorizing requests and interacting with the metastore (database). Upon approval, the application communicates with the Klaw Cluster API application.

Security: The communication between APIs in the governance layer and the cluster layer is highly secure. JWT token-based authentication is used to ensure that no external user can interact directly with it.

User roles and permissions: Every user in Klaw is associated with a role and a set of permissions. Additionally, users can be part of multiple teams and have the flexibility to switch between them.

• Cluster layer: The cluster layer is the second Java application within Klaw. Communication: This layer is a Java application that communicates with the governance layer and Apache Kafka clusters (Kafka, Schema Registry, Kafka Connect).

• User interface switch: By default, users are logged into the AngularJS-based interface. However, they can switch to the React JS interface. Building React assets requires npm, pnpm, and node.

• Metastore Klaw organizes data in the database into three categories:

Administrator data: Comprises users, roles, permissions, teams, tenants, clusters, environments, properties and other product related details.

Core data: Comprises topics, ACLs, schemas and connector configurations.

Requests data: Comprises requests of topics, ACLs, schemas and connectors.

• Cache Klaw stores most frequently queried data in a local cache to for improved performance and user experience. This effectively reduces latency and gives users immediate response from the application. However, this cache is reset whenever changes are requested. Deploying Klaw in different environments like Development, Testing, Acceptance, and Production is essential to streamline the developer experience.

System requirements​

For optimal performance of Klaw, we recommend the following system requirements. These specifications assume a user base of approximately 100 or more, with around 50 or more Apache Kafka clusters and over 10,000 topics.

Core governance application / cluster application

| ------------------------ ------ ------------ | | Deployment environment | RAM | CPU | | ------------------------ ------ ------------ | | Development | 2 GB | Dual-core | | ------------------------ ------ ------------ | | Testing | 2 GB | Dual-core | | ------------------------ ------ ------------ | | Production (HA) | 4 GB | Dual-core | | ------------------------ ------ ------------ |

Database management system​

For the RDBMS, Klaw is compatible with various database management systems such as PostgreSQL®, MySQL, and others.

Note: While the above configurations have been tested and proven to work effectively, there are no guarantees that they will suit every use case. The actual performance and suitability depend on various factors, including the operating system, CPU utilization, and other processes running on the virtual machines.

Deployment model​

To further enhance the system's reliability, deploying the governance application and the Klaw Cluster API application on separate machines is recommended. This setup minimizes the risk of both applications being affected by a single point of failure.

With the understanding of Klaw's working mechanism, let's explore how to deploy Klaw in High Availability production-like environments using the Nginx load balancer.

Nginx load balancer​

In this section, let's explore how to achieve High Availability for Klaw using Nginx as the load balancer. Nginx serves as an efficient HTTP load balancer, distributing traffic across multiple application servers, thereby enhancing the performance, scalability, and reliability of web applications.

Load balancer methods​

Nginx supports three types of load balancing methods:

• Round-robin: Requests are distributed among the application servers in a cyclic manner.
• Least-connected: The next request is directed to the server with the fewest active connections.
• IP-hash: A hash-function based on the client's IP address is used to determine the server to which the next request should be sent.

Detailed information can be found on the official Nginx documentation.

Database authentication in Klaw​

In Klaw, you can configure database-level authentication by setting klaw.login.authentication.type: db in the Klaw Core mode application properties.

With database authentication, Klaw uses the Spring JSESSION ID. When deploying Klaw in HA mode across more than one instance to ensure uninterrupted user access, it's advisable to configure the IP-Hash load balancing method.

Below is an example of Nginx configuration using the IP-Hash method:

Nginx configuration​

upstream klawgcp {
    ip_hash; // load balancing method
    server klawhost1:9097;
    server klawhost2:9097;
  }
  server {
    listen 80;
    listen [::]:80;
    ssl on;

    ssl_certificate     /opt/klaw/certs/service.cert;
    ssl_certificate_key /opt/klaw/certs/service.key;
    ssl_protocols TLSv1.2;
    server_name serverhost;
    location / {
          proxy_pass https://klawgcp;
      }
   }

Using IP-Hash method, sessions are maintained by tracking the client's IP address. Single Sign-On (SSO) authentication in Klaw For SSO authentication, configure Klaw by setting klaw.login.authentication.type: ad in the Klaw Core mode application properties. When SSO is enabled, either Round-Robin or Least-Connected load balancing methods can be used.

Below is an example Nginx configuration using Round-Robin load balancing:

upstream klawgcp {
        round-robin; // load balancing method
        server klawhost1:9097;
        server klawhost2:9097;
    }
    server {
            listen 80;
            listen [::]:80;
            ssl on;

            ssl_certificate     /opt/klaw/certs/service.cert;
            ssl_certificate_key /opt/klaw/certs/service.key;
            ssl_protocols TLSv1.2;
            server_name serverhost;
            location / {
                    proxy_pass https://klawgcp;
                }
    }

Klaw deployment model in High Availability (HA) mode​

In the High Availability mode, Nginx routes the requests to Klaw instances which are connected to a common data store such as Postgres. Each Klaw instance comprises both the governance and Klaw Cluster API applications. You may choose to deploy these applications on separate machines for increased robustness. Below is the deployment model of Klaw in HA mode.

Klaw in HA mode with Nginx Load balancer

Klaw configuration​

While Klaw stores all metadata in a database, most of this data is usually cached for quicker access. Therefore, it's important to reset this cache whenever any changes are made to the configuration of topics, ACLs, etc.

To ensure proper cache reset across Klaw instances, you must configure the following property with the comma-separated list of instance hosts:

klaw.uiapi.servers=https://klawhost1:port,https:klawhost2:port..

This configuration ensures all requests are directed to the various Klaw instances using the Nginx load-balancing configuration.

Other Load Balancers​

Spring Cloud Load Balancer: Allows client-side load balancing. For more information, see the official guide.

Netflix Ribbon: Provides client-side load balancing for Spring Boot applications. See the GitHub repository for more details.

AWS Load Balancer: AWS offers a variety of load balancers based on network, containers, applications, and target groups. Choose the appropriate one based on your requirements. Learn more on the official AWS page.

Klaw project resources​

Project GitHub repository

Git issues

Contribute

Documentation

Community

Docker Klaw Core, Docker Klaw Cluster API

For any questions or discussions, please open an issue on GitHub or participate in our Community forum.

Background​

In today's ever-evolving technological landscape, organizations worldwide have adopted Apache Kafka® for its unparalleled scalability and exceptional features, leading to a paradigm shift in how messaging systems are handled. As more and more organizations adopt Apache Kafka as a vital component of their technology stacks, the need for effective Apache Kafka governance becomes paramount.

A common challenge companies face is the inability to effectively govern and manage access permissions for the numerous topics created, often numbering in the hundreds or even thousands.

Managing Apache Kafka topics poses significant challenges for companies, often leading to struggles executing the task effectively. Several crucial considerations arise when undertaking this endeavor:

• Authorization and ownership: Determining who has the rights to produce and consume messages on a specific topic and establishing clear ownership are crucial aspects to address.
• Configuration backup and restoration: Establishing robust strategies for backing up and restoring Apache Kafka's configuration is essential to ensure prompt recovery in the face of failures or system disruptions.
• Security measures: Ensuring the proper enforcement of security measures through subscriptions and access controls is paramount to safeguarding sensitive data and preventing unauthorized access.
• Topic promotion: Navigating the process of smoothly promoting topics with thoroughly tested configurations from lower environments to production requires careful planning and coordination.
• Enhancing visibility and awareness: Increasing the visibility and awareness of topics within the organization facilitates better collaboration and knowledge sharing among teams.

Managing a small number of Apache Kafka topics may seem straightforward at first. However, as Apache Kafka usage grows and more applications and teams are added, effective and efficient management can become increasingly challenging. This includes the important task of isolating sensitive topics containing GDPR, HIPAA, or other types of sensitive data. It is essential to ensure that only authorized applications can access these topics to maintain data security and regulatory compliance.

Typical challenges​

The current system faces several challenges that impede efficiency and effectiveness, posing risks and hindering productivity. Identifying and addressing these typical challenges can enhance operations and optimize overall performance.

Let's explore the key challenges that need to be addressed:

• Manual activities

  Creating, accessing, and promoting a topic to a higher environment requires extensive email communication, involving approximately 10 emails. This manual approach consumes valuable time and increases potential errors and miscommunication.

• Longer lead time

  The end-to-end communication and execution of the manual activities take approximately 2.5 hours, excluding the time spent waiting for human responses. This long lead time hampers efficiency and slows down the overall workflow.

• Security

  The system faces Role-Based Access Control (RBAC) and data security-related challenges. Ensuring proper access rights and safeguarding sensitive information is crucial for maintaining the integrity and confidentiality of the system.

• Lack of audit

  No proper auditing mechanism is in place to track and monitor requests. This lack of audit trails creates an accountability gap, making identifying and rectifying potential issues or discrepancies challenging.

• Adhoc governance

  The absence of a centralized governance setup adds complexity to the system. All actions are initiated through emails, and the request approver list is maintained in spreadsheets or confluence.

• Release management

  The manual release processes, mainly promoting configurations from one environment to another, introduce a significant risk of errors. These errors can cause system outages, leading to disruptions in service and impacting user experience.

Navigating Apache Kafka system failures​

Failures are an inevitable part of any system, and Apache Kafka applications are no different. These failures can occur at different levels, including client-side issues where event deserialization creates problems for consumers, as well as infrastructure-related issues that result in the loss of ZooKeeper/Apache Kafka configurations.

In the context of infrastructure failures, let's consider a scenario with 1000 topics, each possessing a unique configuration stored in a ZooKeeper. Now, imagine a situation where a disaster strikes, causing the ZooKeeper to crash and be unable to recover. This raises a critical question: how swiftly can you retrieve the information and reboot your server? It's not just about having methods to store information but also about strategically implementing storage and retrieval processes for efficient recovery.

How to mitigate these problems?​

To ensure a fail-safe Apache Kafka system, it is crucial to mitigate these challenges by implementing a robust governance platform. This platform should encompass the following essential capabilities:

• Comprehensive visibility: Providing a holistic view of topics, producers, consumers, schemas, and connectors.
• Secure configuration management: Implementing a four-eye principle-based system for reviews and approvals to maintain the safety and consistency of configurations.
• Notifications: Alerting users about changes to schemas and other configurations.
• Action tracking: Maintaining a detailed record of actions performed by different users for transparent accountability.

Klaw​

Klaw is a comprehensive toolkit designed to address the challenges Apache Kafka developers encounter. It provides solutions for compliance and audit concerns, data security, and release management.

Developed as an open tool, Klaw is committed to contributing to the streaming community by empowering developers to leverage modern governance, control, and security practices while building Apache Kafka applications.

Klaw features​

Klaw is built upon four essential pillars, which serve as the foundation for its capabilities.

Governance - Self-service - Security - Automation

• Governance: Ensures accountability and consistency with approvals, audit trails, naming conventions, clear ownership, and a single source of truth.
• Self-service: Empowers developers with search, customizable dashboard, reporting, and multi-tenancy for efficient resource management and data integrity.
• Security: Offers SSO/AD authentication, RBAC, Apache Kafka Security protocols, audit trails, and multi-tenancy to address diverse security needs.
• Automation: Streamlines tasks like resource provisioning, topic promotion, metadata sync, backup/restore, and notifications for time and cost savings.

Benefits of Klaw​

• Reduced dependency: By reducing dependency on infrastructure teams and empowering developer teams, Klaw enables faster execution and decision-making, saving time, cost, and effort.
• Zero risk of manual errors: With full self-service capabilities, Klaw eliminates the risk of manual errors, ensuring data accuracy and cost savings.
• Continuous integration: Klaw enables continuous integration by promoting topics and schemas to higher environments, ensuring consistency throughout the development pipeline.
• Secure authentication: Klaw authenticates users based on an organization's single sign-on (SSO) or Active Directory (AD), providing a secure access control mechanism.
• Tailored security features: Klaw offers robust security features that cater to the specific needs of different types of organizations, ensuring data protection and compliance.

Klaw project resources​

Project GitHub repository

Git issues

Contribute

Documentation

Community

Docker Klaw Core, Docker Klaw Cluster API

Conclusion​

Organizations are devising unique strategies to tackle the growing complexity of managing an increasing number of Apache Kafka topics, schemas, and associated issues. These strategies involve using tools like Excel, Confluence, and GitOps for partial data maintenance.

For a comprehensive and streamlined solution that simplifies the tasks for both developers and administrators, Klaw stands as a highly effective one-stop solution.