Klyff Platform
Klyff Platform Cloud Edge IoT Gateway Trendz Analytics Mobile Application MQTT Broker
Getting Started Documentation Devices Library Guides Installation Architecture API FAQ
On this page

Cluster setup with Docker Compose

This guide will help you to setup Klyff in cluster mode using Docker Compose tool.

Prerequisites

Klyff Microservices are running in dockerized environment. Before starting please make sure Docker Engine and Docker Compose are installed in your system.

Install Docker:

doc warn icon

Don’t forget to add your linux user to the docker group. See Manage Docker as a non-root user.

Step 1. Pull Klyff CE Images

Make sure your have logged in to docker hub using command line.

1
2
3
4
5
6
7
8

docker pull thingsboard/tb-node:3.8.1
docker pull thingsboard/tb-web-ui:3.8.1
docker pull thingsboard/tb-js-executor:3.8.1
docker pull thingsboard/tb-http-transport:3.8.1
docker pull thingsboard/tb-mqtt-transport:3.8.1
docker pull thingsboard/tb-coap-transport:3.8.1
docker pull thingsboard/tb-lwm2m-transport:3.8.1
docker pull thingsboard/tb-snmp-transport:3.8.1

Step 2. Review the architecture page

Starting Klyff v2.2, it is possible to install Klyff cluster using new microservices architecture and docker containers. See microservices architecture page for more details.

Step 3. Clone Klyff CE repository

1
2

git clone -b release-3.8 https://github.com/thingsboard/thingsboard.git --depth 1
cd thingsboard/docker

Step 4. Configure Klyff database

Before performing initial installation you can configure the type of database to be used with Klyff. In order to set database type change the value of DATABASE variable in .env file:

1

nano .env

to one of the following:

  • postgres - use PostgreSQL database;
  • hybrid - use PostgreSQL for entities database and Cassandra for time series database;

NOTE: According to the database type corresponding docker service will be deployed (see docker-compose.postgres.yml, docker-compose.hybrid.yml for details).

Step 5. Choose Klyff queue service

Klyff is able to use various messaging systems/brokers for storing the messages and communication between Klyff services. How to choose the right queue implementation?

In Memory queue implementation is not suitable for any sort of cluster deployments.

  • Kafka is recommended for production deployments and used by default. This queue is used on the most of Klyff production environments now. It is useful for both on-prem and private cloud deployments. It is also useful if you like to stay independent from your cloud provider. However, some providers also have managed services for Kafka. See AWS MSK for example.

  • RabbitMQ is recommended if you don’t have much load and you already have experience with this messaging system.

  • AWS SQS is a fully managed message queuing service from AWS. Useful if you plan to deploy Klyff on AWS.

  • Google Pub/Sub is a fully managed message queuing service from Google. Useful if you plan to deploy Klyff on Google Cloud.

  • Azure Service Bus is a fully managed message queuing service from Azure. Useful if you plan to deploy Klyff on Azure.

  • Confluent Cloud is a fully managed streaming platform based on Kafka. Useful for a cloud agnostic deployments.

See corresponding architecture page and rule engine page for more details.

Kafka (default, recommended for on-prem, production installations)
AWS SQS (managed service from AWS)
Google Pub/Sub (managed service from Google)
Azure Service Bus (managed service from Azure)
RabbitMQ (for small on-prem installations)
Confluent Cloud (Event Streaming Platform based on Kafka)

Apache Kafka is an open-source stream-processing software platform.

Configure Klyff environment file:

1

nano .env

Check following line:

1

TB_QUEUE_TYPE=kafka

AWS SQS Configuration

To access AWS SQS service, you first need to create an AWS account.

To work with AWS SQS service you will need to create your next credentials using this instruction:

  • Access key ID
  • Secret access key

Configure Klyff environment file:

1

nano .env

Check following line:

1

TB_QUEUE_TYPE=aws-sqs

Configure AWS SQS environment file for Klyff queue service:

1

nano queue-aws-sqs.env

Don’t forget to replace “YOUR_KEY”, “YOUR_SECRET” with your real AWS SQS IAM user credentials and “YOUR_REGION” with your real AWS SQS account region:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32

TB_QUEUE_TYPE=aws-sqs
TB_QUEUE_AWS_SQS_ACCESS_KEY_ID=YOUR_KEY
TB_QUEUE_AWS_SQS_SECRET_ACCESS_KEY=YOUR_SECRET
TB_QUEUE_AWS_SQS_REGION=YOUR_REGION


# These params affect the number of requests per second from each partitions per each queue.
# Number of requests to particular Message Queue is calculated based on the formula:
# ((Number of Rule Engine and Core Queues) * (Number of partitions per Queue) + (Number of transport queues)
#  + (Number of microservices) + (Number of JS executors)) * 1000 / POLL_INTERVAL_MS
# For example, number of requests based on default parameters is:

# Rule Engine queues:
# Main 10 partitions + HighPriority 10 partitions + SequentialByOriginator 10 partitions = 30
# Core queue 10 partitions
# Transport request Queue + response Queue = 2
# Rule Engine Transport notifications Queue + Core Transport notifications Queue = 2
# Total = 44
# Number of requests per second = 44 * 1000 / 25 = 1760 requests

# Based on the use case, you can compromise latency and decrease number of partitions/requests to the queue, if the message load is low.
# By UI set the parameters - interval (1000) and partitions (1) for Rule Engine queues.
# Sample parameters to fit into 10 requests per second on a "monolith" deployment: 

TB_QUEUE_CORE_POLL_INTERVAL_MS=1000
TB_QUEUE_CORE_PARTITIONS=2
TB_QUEUE_RULE_ENGINE_POLL_INTERVAL_MS=1000
TB_QUEUE_TRANSPORT_REQUEST_POLL_INTERVAL_MS=1000
TB_QUEUE_TRANSPORT_RESPONSE_POLL_INTERVAL_MS=1000
TB_QUEUE_TRANSPORT_NOTIFICATIONS_POLL_INTERVAL_MS=1000
TB_QUEUE_VC_INTERVAL_MS=1000
TB_QUEUE_VC_PARTITIONS=1

You can update default Rule Engine queues configuration using UI. More about Klyff Rule Engine queues see in documentation.

Google Pub/Sub Configuration

To access Pub/Sub service, you first need to create an Google cloud account.

To work with Pub/Sub service you will need to create a project using this instruction.

Create service account credentials with the role “Editor” or “Admin” using this instruction, and save json file with your service account credentials step 9 here.

Configure Klyff environment file:

1

nano .env

Check following line:

1

TB_QUEUE_TYPE=pubsub

Configure Pub/Sub environment file for Klyff queue service:

1

nano queue-pubsub.env

Don’t forget to replace “YOUR_PROJECT_ID”, “YOUR_SERVICE_ACCOUNT” with your real Pub/Sub project id, and service account (it is whole data from json file):

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30

TB_QUEUE_TYPE=pubsub
TB_QUEUE_PUBSUB_PROJECT_ID=YOUR_PROJECT_ID
TB_QUEUE_PUBSUB_SERVICE_ACCOUNT=YOUR_SERVICE_ACCOUNT

# These params affect the number of requests per second from each partitions per each queue.
# Number of requests to particular Message Queue is calculated based on the formula:
# ((Number of Rule Engine and Core Queues) * (Number of partitions per Queue) + (Number of transport queues)
#  + (Number of microservices) + (Number of JS executors)) * 1000 / POLL_INTERVAL_MS
# For example, number of requests based on default parameters is:

# Rule Engine queues:
# Main 10 partitions + HighPriority 10 partitions + SequentialByOriginator 10 partitions = 30
# Core queue 10 partitions
# Transport request Queue + response Queue = 2
# Rule Engine Transport notifications Queue + Core Transport notifications Queue = 2
# Total = 44
# Number of requests per second = 44 * 1000 / 25 = 1760 requests

# Based on the use case, you can compromise latency and decrease number of partitions/requests to the queue, if the message load is low.
# By UI set the parameters - interval (1000) and partitions (1) for Rule Engine queues.
# Sample parameters to fit into 10 requests per second on a "monolith" deployment: 

TB_QUEUE_CORE_POLL_INTERVAL_MS=1000
TB_QUEUE_CORE_PARTITIONS=2
TB_QUEUE_RULE_ENGINE_POLL_INTERVAL_MS=1000
TB_QUEUE_TRANSPORT_REQUEST_POLL_INTERVAL_MS=1000
TB_QUEUE_TRANSPORT_RESPONSE_POLL_INTERVAL_MS=1000
TB_QUEUE_TRANSPORT_NOTIFICATIONS_POLL_INTERVAL_MS=1000
TB_QUEUE_VC_INTERVAL_MS=1000
TB_QUEUE_VC_PARTITIONS=1

You can update default Rule Engine queues configuration using UI. More about Klyff Rule Engine queues see in documentation.

Azure Service Bus Configuration

To access Azure Service Bus, you first need to create an Azure account.

To work with Service Bus service you will need to create a Service Bus Namespace using this instruction.

Create Shared Access Signature using this instruction.

Configure Klyff environment file:

1

nano .env

Check following line:

1

TB_QUEUE_TYPE=service-bus

Configure Service Bus environment file for Klyff queue service:

1

nano queue-service-bus.env

Don’t forget to replace “YOUR_NAMESPACE_NAME” with your real Service Bus namespace name, and “YOUR_SAS_KEY_NAME”, “YOUR_SAS_KEY” with your real Service Bus credentials. Note: “YOUR_SAS_KEY_NAME” it is “SAS Policy”, “YOUR_SAS_KEY” it is “SAS Policy Primary Key”:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34

TB_QUEUE_TYPE=service-bus
TB_QUEUE_SERVICE_BUS_NAMESPACE_NAME=YOUR_NAMESPACE_NAME
TB_QUEUE_SERVICE_BUS_SAS_KEY_NAME=YOUR_SAS_KEY_NAME
TB_QUEUE_SERVICE_BUS_SAS_KEY=YOUR_SAS_KEY

# These params affect the number of requests per second from each partitions per each queue.
# Number of requests to particular Message Queue is calculated based on the formula:
# ((Number of Rule Engine and Core Queues) * (Number of partitions per Queue) + (Number of transport queues)
#  + (Number of microservices) + (Number of JS executors)) * 1000 / POLL_INTERVAL_MS
# For example, number of requests based on default parameters is:

# Rule Engine queues:
# Main 10 partitions + HighPriority 10 partitions + SequentialByOriginator 10 partitions = 30
# Core queue 10 partitions
# Transport request Queue + response Queue = 2
# Rule Engine Transport notifications Queue + Core Transport notifications Queue = 2
# Total = 44
# Number of requests per second = 44 * 1000 / 25 = 1760 requests

# Based on the use case, you can compromise latency and decrease number of partitions/requests to the queue, if the message load is low.
# Sample parameters to fit into 10 requests per second on a "monolith" deployment: 

TB_QUEUE_CORE_POLL_INTERVAL_MS=1000
TB_QUEUE_CORE_PARTITIONS=2
TB_QUEUE_RULE_ENGINE_POLL_INTERVAL_MS=1000
TB_QUEUE_RE_MAIN_POLL_INTERVAL_MS=1000
TB_QUEUE_RE_MAIN_PARTITIONS=2
TB_QUEUE_RE_HP_POLL_INTERVAL_MS=1000
TB_QUEUE_RE_HP_PARTITIONS=1
TB_QUEUE_RE_SQ_POLL_INTERVAL_MS=1000
TB_QUEUE_RE_SQ_PARTITIONS=1
TB_QUEUE_TRANSPORT_REQUEST_POLL_INTERVAL_MS=1000
TB_QUEUE_TRANSPORT_RESPONSE_POLL_INTERVAL_MS=1000
TB_QUEU_TRANSPORT_NOTIFICATIONS_POLL_INTERVAL_MS=1000

For installing RabbitMQ use this instruction.

Configure Klyff environment file:

1

nano .env

Check following line:

1

TB_QUEUE_TYPE=rabbitmq

Configure RabbitMQ environment file for Klyff queue service:

1

nano queue-rabbitmq.env

Don’t forget to replace “YOUR_USERNAME” and “YOUR_PASSWORD” with your real user credentials, “localhost” and “5672” with your real RabbitMQ host and port:

1
2
3
4
5

TB_QUEUE_TYPE=rabbitmq
TB_QUEUE_RABBIT_MQ_HOST=localhost
TB_QUEUE_RABBIT_MQ_PORT=5672
TB_QUEUE_RABBIT_MQ_USERNAME=YOUR_USERNAME
TB_QUEUE_RABBIT_MQ_PASSWORD=YOUR_PASSWORD

Confluent Cloud Configuration

To access Confluent Cloud you should first create an account, then create a Kafka cluster and get your API Key.

Configure Klyff environment file:

1

nano .env

Check following line:

1

TB_QUEUE_TYPE=confluent

Configure Confluent Cloud environment file for Klyff queue service:

1

nano queue-confluent-cloud.env

Don’t forget to replace “CLUSTER_API_KEY”, “CLUSTER_API_SECRET” and “confluent.cloud:9092” with your real Confluent Cloud bootstrap servers:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45

TB_QUEUE_TYPE=kafka

TB_KAFKA_SERVERS=confluent.cloud:9092
TB_QUEUE_KAFKA_REPLICATION_FACTOR=3

TB_QUEUE_KAFKA_USE_CONFLUENT_CLOUD=true
TB_QUEUE_KAFKA_CONFLUENT_SSL_ALGORITHM=https
TB_QUEUE_KAFKA_CONFLUENT_SASL_MECHANISM=PLAIN
TB_QUEUE_KAFKA_CONFLUENT_SASL_JAAS_CONFIG='org.apache.kafka.common.security.plain.PlainLoginModule required username="CLUSTER_API_KEY" password="CLUSTER_API_SECRET";'
TB_QUEUE_KAFKA_CONFLUENT_SECURITY_PROTOCOL=SASL_SSL
TB_QUEUE_KAFKA_CONFLUENT_USERNAME=CLUSTER_API_KEY
TB_QUEUE_KAFKA_CONFLUENT_PASSWORD=CLUSTER_API_SECRET

TB_QUEUE_KAFKA_RE_TOPIC_PROPERTIES=retention.ms:604800000;segment.bytes:52428800;retention.bytes:1048576000
TB_QUEUE_KAFKA_CORE_TOPIC_PROPERTIES=retention.ms:604800000;segment.bytes:52428800;retention.bytes:1048576000
TB_QUEUE_KAFKA_TA_TOPIC_PROPERTIES=retention.ms:604800000;segment.bytes:52428800;retention.bytes:1048576000
TB_QUEUE_KAFKA_NOTIFICATIONS_TOPIC_PROPERTIES=retention.ms:604800000;segment.bytes:52428800;retention.bytes:1048576000
TB_QUEUE_KAFKA_JE_TOPIC_PROPERTIES=retention.ms:604800000;segment.bytes:52428800;retention.bytes:104857600

# These params affect the number of requests per second from each partitions per each queue.
# Number of requests to particular Message Queue is calculated based on the formula:
# ((Number of Rule Engine and Core Queues) * (Number of partitions per Queue) + (Number of transport queues)
#  + (Number of microservices) + (Number of JS executors)) * 1000 / POLL_INTERVAL_MS
# For example, number of requests based on default parameters is:

# Rule Engine queues:
# Main 10 partitions + HighPriority 10 partitions + SequentialByOriginator 10 partitions = 30
# Core queue 10 partitions
# Transport request Queue + response Queue = 2
# Rule Engine Transport notifications Queue + Core Transport notifications Queue = 2
# Total = 44
# Number of requests per second = 44 * 1000 / 25 = 1760 requests

# Based on the use case, you can compromise latency and decrease number of partitions/requests to the queue, if the message load is low.
# By UI set the parameters - interval (1000) and partitions (1) for Rule Engine queues.
# Sample parameters to fit into 10 requests per second on a "monolith" deployment: 

TB_QUEUE_CORE_POLL_INTERVAL_MS=1000
TB_QUEUE_CORE_PARTITIONS=2
TB_QUEUE_RULE_ENGINE_POLL_INTERVAL_MS=1000
TB_QUEUE_TRANSPORT_REQUEST_POLL_INTERVAL_MS=1000
TB_QUEUE_TRANSPORT_RESPONSE_POLL_INTERVAL_MS=1000
TB_QUEUE_TRANSPORT_NOTIFICATIONS_POLL_INTERVAL_MS=1000
TB_QUEUE_VC_INTERVAL_MS=1000
TB_QUEUE_VC_PARTITIONS=1

You can update default Rule Engine queues configuration using UI. More about Klyff Rule Engine queues see in documentation.

Step 6. Enable monitoring (optional)

In order to start cluster monitoring - Grafana and Prometheus services, please edit configuration file:

1

nano .env

You’ll need to make sure that MONITORING_ENABLED variable set to true:

1

MONITORING_ENABLED=true

After deployment, you will be able to reach Prometheus at http://localhost:9090 and Grafana at http://localhost:3000 (default login is admin and password foobar).

Step 7. Running

Execute the following command to create log folders for the services and chown of these folders to the docker container users. To be able to change user, chown command is used, which requires sudo permissions (script will request password for a sudo access):

1

./docker-create-log-folders.sh

Execute the following command to run installation:

1

./docker-install-tb.sh --loadDemo

Where:

  • --loadDemo - optional argument. Whether to load additional demo data.

Execute the following command to start services:

1

./docker-start-services.sh

After a while when all services will be successfully started you can open http://{your-host-ip} in you browser (for ex. http://localhost). You should see Klyff login page.

Use the following default credentials:

  • System Administrator: sysadmin@thingsboard.org / sysadmin

If you installed DataBase with demo data (using --loadDemo flag) you can also use the following credentials:

  • Tenant Administrator: tenant@thingsboard.org / tenant
  • Customer User: customer@thingsboard.org / customer

In case of any issues you can examine service logs for errors. For example to see Klyff node logs execute the following command:

1

docker compose logs -f tb-core1 tb-rule-engine1
Doc info icon

If you still rely on Docker Compose as docker-compose (with a hyphen) execute next command:

docker-compose logs -f tb-core1 tb-rule-engine1

Or use the following command to see the state of all the containers:

1

docker compose ps
Doc info icon

If you still rely on Docker Compose as docker-compose (with a hyphen) execute next command:

docker-compose ps

Use the following command to inspect the logs of all running services:

1

docker compose logs -f
Doc info icon

If you still rely on Docker Compose as docker-compose (with a hyphen) execute next command:

docker-compose logs -f

See docker-compose logs command reference for details.

Execute the following command to stop services:

1

./docker-stop-services.sh

Execute the following command to stop and completely remove deployed docker containers:

1

./docker-remove-services.sh

Execute the following command to update particular or all services (pull newer docker image and rebuild container):

1

./docker-update-service.sh [SERVICE...]

Where:

  • [SERVICE...] - list of services to update (defined in docker-compose configurations). If not specified all services will be updated.

Upgrading

In case when database upgrade is needed, edit .env file to set “TB_VERSION” to target version (e.g. set it to 3.8.1 if you are upgrading to the latest). Then, execute the following commands:

1
2
3

./docker-stop-services.sh
./docker-upgrade-tb.sh --fromVersion=[FROM_VERSION]
./docker-start-services.sh

Where FROM_VERSION - from which version upgrade should be started. See Upgrade Instructions for valid fromVersion values. Note, that you have to upgrade versions one by one (for example 3.6.1 -> 3.6.2 -> 3.6.3 etc).

Generate certificate for HTTPS

We using HAproxy for proxying traffic to containers and for web UI by default we using 80 and 443 ports. For using HTTPS with a valid certificate, execute these commands:

1
2

docker exec haproxy-certbot certbot-certonly --domain your_domain --email your_email
docker exec haproxy-certbot haproxy-refresh

NOTE: Valid certificate used only, when you visit web UI by domain in URL. If you using IP address for access to UI, this would be used self-signed certificate.

Next steps