The article describes the development of a functional model and the application of an enterprise email system security subsystem. This addresses the problem of email delivery and ensures the prevention of email loss. The article presents the results of a case study on a company's infrastructure, identifying gaps in the existing system. It discusses the selection of an infrastructure and cloud provider, considering the benefits of containerization and cloud computing. The process of selecting a mail authorization server tailored to the organization's unique email requirements is explored. The article's conclusion underlines the usefulness and significance of the developed method for guarding against email loss in businesses that rely largely on email to effectively communicate with clients and partners.
Email is an important communication tool in
today’s digital world, and reliable email delivery
is essential for the efficient functioning of
businesses and organizations. However,
periodic outages of Internet DNS services and
interruptions in the operation of email service
providers emphasize the need for organizations
to have their email delivery infrastructure [
The loss of emails while they are being sent
from containerized environments is a serious
problem that can significantly impact
communication efficiency and business
processes [
This problem has important implications for
many organizations that rely on email to
communicate with customers, partners, and
internal departments. Losing emails can lead to
data loss, missed deadlines, poor stakeholder
communication, and ultimately financial loss [
However, the solution to this problem is not so simple. Developing a mechanism for creating a mail queue in an application is not cheap and involves certain technical challenges. In addition, containerized environments may contain applications whose source code is not accessible, which complicates the process of identifying the causes of lost emails. The goal of this study is to come up with a method to stop emails from being lost when sent from containerized environments and to address this issue. We propose to conduct research aimed at developing an effective mechanism that will reduce the risk of email loss. The study will analyze the existing problems of email delivery, technologies, and methods of preventing email loss, taking into account the technical limitations of containerization. 1.1.
Emails have become an effective initial vector of infection because virtually every company uses email, and employees receive a large volume of emails every day. Because of the overwhelming volume of emails, employees have very little time to review and analyze each one, which might give them a false sense of security. Cybercriminals are effectively exploiting this situation by organizing phishing attacks, which are becoming even more widespread and sophisticated with the advent of cloud-based email services. Consequently, email security is becoming a critical task for businesses and employees to avoid the consequences of these evolving threats.
Simple Mail Transfer Protocol (SMTP) [
In response to these shortcomings, Postfix was developed as an alternative to Sendmail. Postfix is designed to address many of the security concerns associated with Sendmail. Additionally, Postfix simplifies the management of email server installations by employing a straightforward approach. Administrators can conveniently handle Postfix through two configuration files, reducing complexity.
One of the standout features of Postfix is its ability to function effectively even in demanding situations. It is not uncommon for software systems, including email servers, to face unexpected conditions such as running out of memory or disk space. Postfix, however, excels at detecting and handling such conditions without worsening the problem. This robustness ensures stable and reliable email operations, even during challenging circumstances.
Of course, like any system that started a long time ago, the email system has evolved alongside and in parallel with the development of the entire Internet.
So, just as the infrastructure of enterprises has been changing first towards hydride and then towards fully cloud-based, the infrastructure also needs to evolve. As of today, more and more applications are migrating to Kubernetes clusters hosted in different cloud environments, which means that not only the applications themselves have to be redesigned for such a containerized environment, but also all the accompanying software.
The main purpose of e-mail is to exchange messages (information that is delivered asynchronously). The main messaging protocol is SMTP, which aims to ensure reliable and efficient exchange of letters.
The SMTP protocol is used to send all email
in the world. According to official
documentation [
A visual representation of the protocol’s scheme is provided in Fig. 1. The network on which SMTP is implemented consists of hosts that communicate using the TCP protocol and operate on the public Internet, an isolated internal TCP/IP network, or other local or wide area networks using another transport layer protocol other than TCP. With SMTP, a process can transmit electronic messages to another process on the same network or a different network using a relay or gateway that is accessible to both networks.
Thus, an electronic message can pass
through several intermediate relays or
gateways on its way from the sender to the
final recipient. This ensures efficient routing
and delivery of emails to their destination [
Computer programs that allow you to create, send, receive, and view emails are called email clients. Users can access email using an interface made by an email client.
Both locally on a user’s device and remotely on a web server are capable of running email clients. For the first case, examples include Microsoft Outlook, Apple Mail, and Mozilla Thunderbird—these programs are installed directly on the user’s computer or mobile device and provide the ability to manage email without a direct connection to the Internet.
In the case of web servers, examples of email clients are Google Gmail and Yahoo! Mail—these services are provided in the cloud and are accessible to users through a web browser. In particular, a user can access his or her email from any device with an Internet connection without having to install separate software on each device.
The Simple Messaging Protocol system consists of many different elements, the main elements of the system are shown in Fig. 2. Separate operations that are in charge of sending and receiving mail are necessary for an exchange system to function successfully.
The incoming mail agent and the outbound mail agent must typically run on separate servers in large commercial enterprises.
To send messages to recipients and to receive new messages from the user’s mailbox, the mail client creates two connections: one to the outgoing mail server and one to the incoming mail server.
The email client connects to the incoming mail server and the outgoing mail server using various interfaces. For communicating with particular mail agents that are used to process incoming and outgoing emails, these can be distinct protocols and ports.
This separation of functionality between
servers and different interfaces allows for
efficient and reliable email management and
ensures optimal interaction with email clients.
The outgoing mail server serves several
important functions:
1. Message Queue: It receives messages
from mail clients and places them in a
queue for further delivery.
2. Routing: The outgoing mail server
determines the appropriate incoming
mail server for each recipient and
transfers the messages to that server for
final delivery.
3. Client-Server Interaction: Acting as a
server, it interacts with mail clients
when receiving messages, but functions
as a client when relaying messages to
incoming mail servers.
4. Connection Initiation: The outgoing mail
server ensures that connections are
always initiated by email clients.
5. Bounce Messages: If the outgoing mail
server is unable to deliver a message, it
sends a bounce message back to the user
who sent the original message.
6. Sender Information: It adds information
about the sender, but does not alter the
content of the message.
7. Authentication: Before accepting a
message, the outgoing mail server
typically authenticates the user, often
requiring a username and password [
In some cases, the inbound mail server may
support an alternative authentication method
such as OAuth. This allows the email client to
obtain a special access token that allows access
to only certain functions or messages within a
limited scope. This approach gives users more
control over which applications or services
have access to their mailboxes [
After studying the principles of exchange, we move on to study the problems of secure mail sending.
An incoming mail server operates by
listening for connections from outgoing mail
servers. When an outgoing mail server
connects to send a message, the incoming mail
server records the message and session details,
including the sender’s IP address [
Incoming messages can be rejected by an of Secure Mail Sending
email server for various reasons, such as:
• Non-existent email recipient. Classifying messages as spam based on their
• Full mailbox of the recipient. content and origin is an important task in spam
• Message size exceeding the server’s limit. filtering systems. Probabilistic classifiers, such
• Unreliable or inauthentic sender. as naive Bayesian spam filtering [
Once the incoming mail server receives a necessary to convert the continuous message, it takes charge of its subsequent probability into a binary value. If the delivery. If there are any issues with delivery, probability exceeds a certain threshold, the such as the need for forwarding, the incoming message is rejected as spam. This approach mail server informs the message sender. allows you to divide messages into two
After concluding the session with the categories: spam and non-spam (legitimate
outgoing mail server, the incoming mail server messages).
includes the gathered session details in the When classifying spam, it is important to
received message. It then assesses the message ensure that the false positive rate (classifying
to determine if it is spam or legitimate. legitimate messages as spam) is close to zero.
Depending on this assessment, the message This helps to avoid users missing important
may be: messages. On the other hand, the false negative
• Delivered to the recipient’s inbox if it is rate (classifying spam as legitimate messages)
deemed legitimate and does not contain can be somewhat higher to prevent spam from
any spam indicators. getting into the inbox [
For authorization, the email client provides identification data, such as an email address and password. The inbound mail server verifies this information to ensure that it corresponds to a valid user and matches the data in the database correctly. restrictions and spam mailing lists. Preventing such attacks and protecting emails are becoming important tasks to ensure the safety and reliability of users.
Modern email protocols are designed to
ensure confidentiality, authenticity, and
integrity during message transmission without
the need for user intervention. The use of
Transport Layer Security (TLS) in IMAP, POP,
and SMTP protocols allows you to ensure the
confidentiality of messages even when an
attacker acts in the middle, trying to capture
information. Additionally, the DomainKeys
Identified Mail (DKIM) mechanism along with
domain name system security extensions
guarantees the authenticity and integrity of
messages [
Support for TLS is common, but certificate
validation is not always used, which can
compromise protection against active attacks
by an attacker [
The case study enterprise, from the very beginning of its application development, had a distributed infrastructure, and applications were hosted on physical servers in various data centers. With the development of cloud computing, application orchestration, and containerization technologies, the company began moving its infrastructure to containers hosted by cloud providers.
At present, the company has not implemented a mechanism for storing and processing e-mail, which in turn leads to its loss and potential interception by a third party.
In this case, the loss or transfer of information or the contents of the emails is critical, as the emails contain information about orders placed and other sensitive information.
We have a situation in which the main applications have already been moved to the cloud, but emails are sent directly to recipients without proper protection and minimal operations to hide the content of emails. The main problem, of course, is the loss of emails in the chain between the application and the recipient. The main task is to develop a clusterembedded solution that combines encryption, message queuing, and logging of the number of sent and unsent messages.
Since there are a large number of applications in the infrastructure, auditing how many users have sent emails is also a critical factor. At the moment, an end user who sends only a few emails and a user who sends hundreds of thousands of emails pay the same because there is currently no control mechanism.
A container is a standardized unit of software
that stores code along with all its dependencies
and allows it to be easily ported and run in
different environments standardization by
including dependencies means that you will get
the same behavior wherever you run it [
Transform your monolithic code into lightweight modules to enhance manageability and connectivity. By doing so, you can avoid the risk of one small module breaking your entire program. This approach grants you more precise control over your code, but it does introduce multiple moving parts to your platform.
However, managing numerous moving parts can be challenging. When one container connects to another, it becomes essential to remember to update each one for platform stability. With dozens of containers, this can lead to code management complexities.
Kubernetes greatly simplifies the process of
deploying containers and helps to reduce a
large number of costs associated with it.
Thanks to it, deployment periods can be
reduced from a full day to a much shorter time,
as Kubernetes automates code compilation,
testing, and checking for updates to all
services. Previously, there were several
automation tools to help with the deployment
process, but they were mostly designed for
monolithic architectures, while Kubernetes
provides more efficient and scalable container
management. Orchestration and automation
offer solutions to the challenges encountered
when deploying infrastructure manually. With
these processes, you can eliminate the risk of
forgetting files, ensure updates are applied
across all servers, and easily undo changes as
needed [
One distinguishing factor between Kubernetes and other deployment options is the continuous nature of container deployment. There’s no need to wait for sequential compilation and deployment of binaries. Instead, Kubernetes constantly incorporates new changes into containers and deploys them in the background. This enables swift code deployment without disrupting performance during specific periods. With Kubernetes and containers, there’s no need for developer intervention to manage services that require frequent updates.
Kubernetes and containers are well-suited for cloud environments due to their portability and lightweight nature. They can be deployed on cloud platforms as well as local servers, making them versatile for a multi-cloud strategy across various service providers. Containers offer an attractive option for reducing risk when implementing a microservice architecture in cloud computing.
When considering the best cloud hosting service out of the three options (AWS, Azure, GCP), it is important to analyze the benefits and drawbacks of each. AWS, being the leader in the cloud hosting market, should be given priority.
AWS offers three container environments: ECS, EKS, and Fargate. If developers have limited experience with containers and already use AWS to host their services, ECS is the recommended option. This “containers-as-aservice” solution automates deployments directly in the cloud using Amazon AWS CloudFormation, making it a great starting point to determine if containers are suitable for an organization.
For a more comprehensive use of Kubernetes and containers, Amazon EKS is a suitable choice. EKS can move an existing onpremises Kubernetes deployment to the cloud. Advantages and surveys show that EKS is positioned to become the most popular container management method, with 63% of container users surveyed by Kubernetes preferring AWS.
AWS Fargate, Amazon’s latest release for container users, enables container deployment without the need to manage servers or clusters. Fargate also works with AWS EKS, providing multiple options and combinations based on individual needs.
Hosted Kubernetes with AWS is particularly
attractive for developers who are new to the
container environment. It serves as an
excellent starting point to experiment with
containers and gauge their compatibility with
the development environment. However, EKS
is considered difficult to set up and requires
technical experience with containers. On the
plus side, it offers full scalability and
configurability, empowering companies to
control Kubernetes and its integration with
local development processes [
Let’s now analyze the advantages and disadvantages of Microsoft Azure. IT professionals primarily working with the Windows software environment will find deploying to Azure intuitive and straightforward. Although Azure is a relatively new container service, launched in 2015, Microsoft continues to enhance its offerings.
While Microsoft is known for its Windows operating system, Azure can also work with various Linux and Unix distributions. This means that it is not limited to Windows-only applications, but it does have limitations regarding hybrid container support, unlike AWS, which does support hybrid deployments.
Azure introduced Azure Kubernetes Service (AKS) in October 2017, a service similar to AWS EKS. Deploying AKS on an Azure virtual machine has the added advantage of being free, with payment only required for the resources used on Azure’s virtual machines.
The main disadvantage of Azure is that while AKS was introduced before AWS EKS, Kubernetes is more adapted to AWS and GCP. As a result, Azure lags behind both AWS and Google’s Kubernetes engine when it comes to updating to the latest versions of Kubernetes. However, if Azure is already a part of your existing architecture and you need to implement containers, it simplifies deployment and provides detailed analytics to assess if the platform meets your needs. Additionally, Azure offers a competing service to AWS ECS called Service Fabric, which is also worth considering.
Now, let’s analyze the advantages and disadvantages of Google Cloud Platform (GCP). Given that Google is the original creator of Kubernetes, working with GCP provides several advantages. Any new versions and deployments are immediately available, while other platforms may have a time lag. Moreover, GCP offers excellent opportunities for working with big data, machine learning, and Artificial Intelligence (AI) technologies.
However, GCP faces challenges as it is not as popular as AWS or Azure in terms of infrastructure as a service solution. GCP lacks the small business cloud offerings that make AWS and Azure more appealing for corporate integration into internal networks. Additionally, GCP does not have features like Active Directory integration (Azure) or Identity and Access Management (AWS).
Google promotes its platform as the ideal choice for working with the DevOps methodology. DevOps teams consist of specialists who simultaneously work on automating deployment and application development tasks. For such teams, GCP offers advantages for automating deployments.
Comparing costs becomes difficult due to
the pay-as-you-go pricing policies of these
platforms. The price an individual developer
experimenting with Kubernetes and
containers pays will not be the same as what an
enterprise requires for powerful computing
resources. Costs also depend on the resources
utilized, with each platform having a minimum
cluster allocation [
Ultimately, there is no clear-cut answer regarding the best service to use with Kubernetes. The choice depends on the specific project requirements. If you are already experienced with Azure or AWS, it is advisable to stick with the same platform. However, if you need to work with AI and machine learning, GCP offers an attractive and costeffective solution. 2.3.
Organizations have different email needs. While some businesses only need a way to send straightforward marketing newsletters, others need a more sophisticated email infrastructure for high-volume and transactional uses.
Let’s now examine the technical details of SendGrid and Mailgun, the two top email platforms. Both systems provide great ways to communicate large amounts of data, complete transactions, and guarantee first-rate delivery quality.
Being the “Email Service for Developers,” Mailgun offers robust APIs that make sending, receiving, and tracking emails simple. Mailgun provides a sophisticated API, Mailjet for marketing, email verification, bulk sending, delivery issue prediction, and other services to over 200,000 enterprises with a 99.99% uptime guarantee.
Let’s now evaluate SendGrid’s advantages and disadvantages. SendGrid, one of the top transactional email systems on the market, also provides email marketing capabilities including user-friendly registration forms and design templates. SendGrid, which has more than 80,000 customers’ trust, enables the safe delivery of more than 70 billion emails each month.
Since its founding in 2009, SendGrid has
maintained an advantage in transactional
email services. They stand out from the
competition because of their emphasis on
availability, scalability, and considerable email
experience. Clients that use SendGrid’s
services enjoy an average delivery rate of 97%,
exceeding the sector average of 85%,
according to the company’s data [
To counter phishing and spoofing attempts, SendGrid provides sophisticated authentication with SPF and DKIM. Furthermore, they include AI-based technologies like their adaptive communication engine, which makes use of AI to improve availability and react to changing ISP requirements.
In addition, Mailgun offers proactive email monitoring, continuing advising, managed delivery services with API support, and management of IP and domain reputation.
A 97% average delivery success rate is what Mailgun advertises. Additionally, they tout an industry-lower bounce rate of 0.4% on average, compared to the industry standard of 2%.
Email security helps protect organizations and recipients from data breaches and other threats.
Use secure login authentication methods.
The traditional use of just a password and login to authenticate a user is considered outdated, but many people still use it. Modern authentication methods require two or more factors to verify a user’s identity.
To use Two-Factor Authentication (2FA), users must combine three different types of authentication factors: knowledge (something you know, like a PIN or password), possession (something you have, like a debit card or cell phone), and physical uniqueness (something that only you have, like a fingerprint). For instance, you have to show your debit card and enter your PIN to use an ATM (possession factor and knowledge component, respectively). Temporary one-time passwords or SMS are frequently used for this kind of authentication.
Users can access numerous applications
using Single Sign-On (SSO) after authenticating
their identities with an identity provider. SSO
and 2FA are frequently coupled. Centralized
access with SSO lessens the possibility of
account intrusion, streamlines the login
procedure, and aids in enforcing strong
passwords [
Application Programming Interfaces (APIs) are unique codes that identify and authenticate users. They provide a more secure alternative to logging in with a username and password. Developers commonly use APIs to control access to services like email APIs.
When configuring API keys, different permission levels can be assigned to restrict access to specific parts of an account.
Sender authentication protocols are essential for proving the authenticity of the sender and preventing spam or spoofing. The following authentication protocols should be used: • SPF: Identifies mail servers allowed to send emails from a specific domain. • DKIM: Verifies that the sender is responsible for the email’s content and domain. • Domain-based Message Authentication, Reporting & Conformance (DMARC): Specifies how to handle emails not authenticated by SPF or DKIM.
Using an SMTP server is crucial for email delivery. The SMTP server processes and sends emails to the recipient’s inbox. It also verifies the email’s origin and protects recipients from illegitimate senders.
SMTP authentication enhances email security by requiring a login using a supported authentication mechanism.
Data protection is vital, especially for
sensitive information. Encryption, such as TLS,
safeguards email content during transmission.
Mail providers prioritize TLS-encrypted
connections when delivering emails,
preventing unauthorized access to email
content [
To enhance data security, it is important to restrict access to data, especially recipient data, to only those employees who require it. This minimizes the risk of email leaks becoming more serious if the data is compromised. Here are some ways to achieve this: • Limit the number of users to only those employees who regularly need access to your ESP. Regularly review the user list to ensure there are no inactive users. • Implement different levels of permissions based on each user’s role. This allows team members to access the necessary features for their functions while safeguarding sensitive customer data.
SPF is an email authentication method that
verifies which mail servers are authorized to
send emails from a specific domain [
Cisco and Yahoo jointly created the
encryption method known as DKIM. Its
function is to give senders the option to “sign”
their messages, enabling receivers to confirm
the legitimacy and authorization of the domain
that sent the email [
DKIM is a straightforward method for email authentication that confirms the sender is in control of the domain and responsible for the email’s content. There are two stages to the DKIM process: • A private key is added by the sender to their email servers, and this key is used to sign the email. • The receiving server verifies the sender’s signature using the public key that is kept in the text record dkimselector._domainkey.domain.com.
DKIM is essential for establishing the sender’s legitimacy, promoting confidence, and lowering the dangers of spam and phishing assaults. Inbox providers may block emails if DKIM is not appropriately implemented, preventing them from getting to the intended recipients. While a few blocked messages might not seem like much, they can have a big impact on businesses.
There are numerous internet resources for checking DKIM. A DKIM analyzer or DKIM validation tool can assist in making sure that your DKIM record has been published correctly. Before putting any modifications to SPF or DKIM data into effect, they must be verified.
Domain-based Message Authentication, Reporting & Conformance
SPF and DKIM are two methods used by the DMARC protocol to confirm the legitimacy of an email message.
Internet Service Providers (ISPs) can successfully resist fraudulent email tactics like domain spoofing, which try to trick recipients and steal their personal information, by using DMARC records.
DMARC enables email senders to declare what to do with messages that are unable to authenticate with SPF or DKIM. Senders can either direct these emails to a garbage folder or completely block them. With fewer false positives because of this proactive method, carriers may more precisely detect spammers and safeguard customers’ inboxes. Additionally, it produces thorough authentication reports for increased market transparency.
Implementing DMARC is highly recommended for the following reasons:
Reputation: By stopping unauthorized parties from sending emails from a domain that is not their own, publishing a DMARC record protects your brand. Sometimes, improving your reputation simply by releasing a DMARC record is enough.
Visibility: By identifying the senders who are using your domain to send emails, DMARC reports give you important information about your email program.
Security: DMARC aids in establishing a uniform protocol for processing unauthenticated emails across the email community. The entire email ecosystem’s security and dependability are strengthened by this group effort.
A crucial addition to email authentication, DMARC data show how email senders and ISPs work together to secure the email channel.
Transport Layer Security
A technology called TLS protects digital
communications between two parties. It
makes sure that no communication may be
intercepted, eavesdropped on, or altered while
data is transmitted between a server and a
client [
The Secure Sockets Layer (SSL) protocol was replaced by the TLS protocol, which is now officially deprecated and regarded as insecure. It is strongly advised against using any SSL version. Additionally, the IETF declared in RFC 8996 that TLS versions 1.1 and 1.0 are no longer supported and should not be utilized. The NCSC advises upgrading TLS 1.1 or 1.0based government systems to either TLS 1.3 or 1.2.
To improve data security, TLS provides three main services: Authentication: This enables each side to verify the other party’s identity throughout the communication.
Data between the user agent and the server is transmitted encrypted to prevent reading or decryption by unauthorized parties.
Integrity: TLS protects data against loss, damage, manipulation, or fabrication by ensuring that it is unchanged and unaltered throughout encryption, transmission, and decryption.
The handshake phase of a TLS connection is where the client and server create a shared secret and talk about crucial factors, including cipher settings, to ensure a secure communication channel.
According to the study of the requirements for the design of the security subsystem, we have a problem with email delivery. It has been determined that emails may not reach the final recipient for the following reasons: • SendGrid may become unavailable. • The SendGrid API key may expire or be incorrectly configured. • IP address restrictions may be configured incorrectly.
That’s why it was decided to add a mail intermediary to this scheme, which will make a mail queue and prevent mail loss. Since we have a multi-cluster environment, we need to take care of working in this scenario. So, each cluster has two redundant Postfix applications running synchronously and on different feeds. On each cluster, Postfix will have separate settings and its own SendGrid API key.
After researching the available software, the following resources were selected: • Azure cluster (AKS). • Postfix as a mail transfer agent. • Fluent-bit for collecting logs from other containers. • Grafana-Loki-SQL database for logging processing. • Azure managed disks for persistent storage to keep the mail queue even when something happens to the deployment.
Develop a system that will store the mail sent by the application, check if the mail is possible, and only then send the mail. You also need to develop a system for monitoring the number of sent emails. We develop Helm configuration files for deploying infrastructure in cluster environments. So, in one cluster we have: • Two Postfix pods that include two containers with Fluent-bit and Postfix. • Two PersistentVolumeClaim. • Two ReplicaSets. • One ConfigMap for FluentBit. • One ConfigMap for Postfix. • One ServiceAccount. • One Service. • One Secret.
The Postfix application was chosen as the basis of the email system for the project. The main problem with Postfix, in our case, is that its development began a long time ago and was not designed for the modern infrastructure of Internet technologies. Postfix has proven to be a simple and reliable system for sending and queuing mail, but it was not designed for containerized environments and has a log system that is quite difficult to work with. For optimal work with logs and system monitoring in general, it was decided to create an application for parsing and transforming logs].
As an alternative to syslog (which is still the default), Postfix has its logging system. Postfix version 3.4 or later supports this. When Postfix is operating in a container, logging to stdout is advantageous since it removes the syslogd dependence.
Python, a general-purpose high-level programming language, was used to create the solution. Its design philosophy places a strong emphasis on the readability of the code by using substantial indentation.
The most recent Alpine base container with the parser and Postfix was chosen as the foundation for the container with the parser. Alpine is currently the standard for using it as the basis for your applications. It is constantly updated, constantly receives modern solutions to security problems, has basic functionality, and, despite all of the above, has a fairly small size of ~5Mb.
However, for Postfix to work properly and use all the parser’s functions, you need to install the software. For the parser to work, you must first install Python from the repository, the pyodbc library, msodbcsql, mssql-tool. For Postfix to work with secure authorization, you need to install Postfix and the packages postfixpcre, libsasl, cyrus-sasl. In addition to all these packages, you need to install some additional ones.
For Postfix authorization to work properly, you need to create a system user and add information about it to the Postfix configuration files.
The final step in the Dockerfile is to run the script for initial setup and start the parser.
The functions of the initial startup script are to configure Postfix parameters and then start the parser. The shell used is ash, which is a simplified version of the shell from the Alpine Linux developers. The parameters used for configuration are taken from environment variables, which is a safe approach.
We have also implemented functionality when it is necessary to transfer non-standard settings when deploying a solution. For this purpose, there is a separate environment variable POSTFIX_CUSTOM_CONFIG, to which non-standard parameters are passed through the “;” symbol.
Today, many cloud service providers provide services for renting their computing capabilities and using Kubernetes clusters based on their equipment. In this case, the architecture of the solution is based on the AKS from Microsoft. Despite this, the development of automated templates based on Helm Charts is not dependent on service providers in most cases and is universal, that is, it can be used on almost any environment, including a local computer.
Helm is a Kubernetes package manager. Helm is used to build “charts”, which are packages of Kubernetes resources used to deploy applications to a cluster. There are different types of storage and distribution of these packages: they can be stored in archives or just directories with files, distributed through private or public repositories.
The main part of any Helm chart is the Deployment type, which describes what the Pod will consist of, what containers should be included in it, what files should be attached to it, and what environment variables should be in the containers.
In this case, the Deployment consists of a Pod that includes two containers: Fluent-Bit of the latest version and the container with the parser and Postfix. The container with FluentBit contains log storage and a configuration file. A persistent memory disk is attached to the container with Postfix to store logs in case of a cluster or container reboot. In Deployment, checks are additionally configured to see if the container has been successfully powered on and if it is still available.
If you need to have multiple copies of Postfix on different host machines for greater reliability, Deployment is configured to prevent the installation of two copies on one host. Also, environment variables that are stored in the ConfigMap are passed to the container. The ConfigMap.yaml file describes two ConfigMap resources: the first one stores environment variables that are passed to the container from Postfix and the parser, and the second one contains configuration files for Fluent-Bit.
The FluentBit settings describe where to get the information, how to process it, how to filter only the necessary data, and where to send the generated output information. It also contains information on how to work with the information we receive from the developed parser.
To store passwords, certificates, etc., the Secret type is used, which stores information in a hashed form. It’s worth noting that it’s exactly hashed, not encrypted. In the Secret of our Helm Chart, we store the password to the MSSQL database and the key to the SMTP service provider. This is to ensure that other services in the cluster, namely applications that send mail, have a network connection to Postfix, you need to create a Service dash resource. It specifies the port-forwarding parameters. that contains The name, version, and version of the program, as well as a brief description of the Helm Chart, are all included in the Chart.yaml file, which also contains some basic information about the Helm Chart itself.
Variables can be passed to Helm Chart using various methods, but the best practice is to store them in a YAML file and keep them in a version-controlled git repository to prevent configuration drift.
It is recommended to deploy one Postfix Deployment per cluster but with a different number of Postfix replicas, at least two.
At this point, the development of the template for automatic deployment of the solution can be considered completed. Further, the received files will be automatically processed by the helm utility, which is part of the Azure DevOps Pipelines portal plugins. After the system has been successfully developed, and the operability of several systems has been constructed, it is necessary to have complete information about the operation of these systems, error visualization, and complete statistics on successfully delivered mail.
There are many systems for monitoring, but the choice was made to use free open-source solutions—Grafana and Grafana Loki.
Grafana Loki is an aggregator of log records, in this case, received from FluentBit.
Grafana is a modern data visualizer that can work with almost any data source. We need to visualize data from Grafana Loki and MSSQL.
Several deployment options are available for Grafana and Grafana Loki, including installation as a service, operation in a container, and deployment in Kubernetes. We chose to use Kubernetes since it is currently the most cutting-edge method of deploying Grafana.
With the help of the IDE for Kubernetes— Lens, we will deploy the above services with standard settings, but add a certificate for secure connection to web resources.
In the Grafana data sources settings, we configure Grafana Loki and MSSQL.
Grafana has a fairly wide range of tools for displaying information, which means that operators will be able to get all the necessary data in the form they need. Monitoring systems are installed on the cluster environment and dashboards are developed for easy access to the necessary information.
The created system to prevent the loss of emails is relevant and important for organizations that rely on email for effective communication with customers and partners.
The efficiency of communication operations can be greatly increased and the risk of email loss can be dramatically decreased by implementing the security subsystem that has been created for a corporate email system based on containerized environments.
Taking into account the technical limitations of containerization and choosing the best technologies (for example, using Postfix, SendGrid, Microsoft Azure, and AKS) helps ensure the stability and reliability of the system.
The deployed data monitoring and visualization system allows us to quickly track mail-sending processes and respond to unforeseen events promptly.
The design of the email security subsystem demonstrates an increased level of reliability and visibility of email services in the enterprise, which can help improve business processes and mutual understanding with stakeholders.
Research and development of an email delivery system is a complex task, but given the importance of electronic communication in the modern world, investments in the development of such a system can be justified and bring positive results for organizations.
In general, the development and implementation of an email security subsystem for an enterprise is a step forward in ensuring reliable and efficient communication, which will help improve interaction with customers and partners and reduce the risks of financial losses associated with data loss and delays in communication processes.