Deploying Apica Ascent PaaS on Kubernetes
This page describes the Apica Ascent deployment on Kubernetes cluster using HELM 3 charts.
1 - Prerequisites
Kubernetes 1.18, 1.19 or 1.20
Helm 3.2.0+
Dynamic PV provisioner support in the underlying infrastructure
ReadWriteMany volumes for deployment scaling
Apica Ascent K8S components are made available as helm charts.
1.1 Add Apica Ascent helm repository
1.1.0 Adding Apica Ascent's helm repository to your HELM repositories
The HELM repository will be named apica-repo
. For installing charts from this repository please make sure to use the repository name as the prefix e.g.
helm install <deployment_name> apica-repo/<chart_name>
You can now run helm search repo apica-repo
to see the available helm charts
1.1.1 Update your HELM chart
If you already added Apica Ascent's HELM repository in the past, you can get updated software releases using helm repo update
1.2 Create namespace where Apica Ascent will be deployed
NOTE: Namespace name cannot be more than 15 characters in length
This will create a namespace apica-ascent
where we will deploy the Apica Ascent Log Insights stack.
If you choose a different name for the namespace, please remember to use the same namespace for the remainder of the steps
1.3 Prepare your Values YAML file
Sample YAML files for small, medium, large cluster configurations can be downloaded at the following links.
These YAML files can be used for deployment with -f parameter as shown below in the description.
Please refer to Section 3.10 for sizing your Apica Ascent cluster as specified in these YAML file Latest image tags.
2. Install Apica Ascent
This will install Apica Ascent and expose the Apica Ascent services and UI on the ingress IP. If you plan to use an AWS S3 bucket, please refer to section 3.2 before running this step. Please refer to Section 3.4 for details about storage class. Service ports are described in the Port details section. You should now be able to go to http://ingress-ip/
The default login and password to use is flash-admin@foo.com
and flash-password
. You can change these in the UI once logged in. HELM chart can override the default admin settings as well. See section 3.7 on customizing the admin settings
Apica Ascent server provides Ingest, log tailing, data indexing, query, and search capabilities. Besides the web-based UI, Apica Ascent also offers apicactl, Apica CLI for accessing the above features.
3 Customizing the deployment
3.1 Enabling https for the UI
You should now be able to login to Apica Ascent UI at your domain using https://ascent.my-domain.com
that you set in the ingress after you have updated your DNS server to point to the Ingress controller service IP
The default login and password to use is flash-admin@foo.com
and flash-password
. You can change these in the UI once logged in.
The ascent.my-domain.com
also fronts all the Apica Ascent service ports as described in the port details section.
HELM Option | Description | Defaults |
---|---|---|
| DNS domain where the Apica Ascent service will be running. This is required for HTTPS | No default |
| Enable the ingress controller to front HTTPS for services | false |
| Specify if a default certificate is enabled for the ingress gateway | false |
| Specify the name of a TLS Secret for the ingress gateway. If this is not specified, a secret is automatically generated of option |
3.1.1 Passing an ingress secret
If you want to pass your own ingress secret, you can do so when installing the HELM chart
3.2 Using an AWS S3 bucket
Depending on your requirements, you may want to host your storage in your own K8S cluster or create a bucket in a cloud provider like AWS.
Please note that cloud providers may charge data transfer costs between regions. It is important that the Apica Ascent cluster be deployed in the same region where the S3 bucket is hosted
3.2.1 Create an access/secret key pair for creating and managing your bucket
Go to AWS IAM console and create an access key and secret key that can be used to create your bucket and manage access to the bucket for writing and reading your log files
3.2.2 Deploy the Apica Ascent helm in gateway mode
Make sure to pass your AWS_ACCESS_KEY_ID
and AWS_SECRET_ACCESS_KEY
and give a bucket name. The S3 gateway acts as a caching gateway and helps reduce API costs.
Create a bucket in AWS s3 with a unique bucket name in the region where you plan to host the deployment.
You will need to create the S3 bucket manually along with access and secret keys to access the bucket. Check to make sure the access and secret key work with the newly created bucket.
Once the bucket is created and access/secret is verified, provide the bucket name and access credentials in the step below.
Additionally, provide a valid amazon service endpoint for s3 else the config wll default to https://s3.us-east-1.amazonaws.com
HELM Option | Description | Defaults |
---|---|---|
| This helm option specifies the supported cloudProvider that is hosting the S3 compatible bucket. Right now only | aws |
| Name of the S3 bucket in AWS | logiq |
| S3 Service endpoint : https://s3.**<region>**.amazonaws.com | |
| AWS Access key for accessing the bucket | No default |
| AWS Secret key for accessing the bucket | No default |
| AWS Region where the bucket is hosted | us-east-1 |
S3 providers may have restrictions on bucket names for e.g. AWS S3 bucket names are globally unique.
3.3 Install Apica Ascent server certificates and Client CA [OPTIONAL]
[OPTIONAL]
Apica Ascent supports TLS for all ingest. We also enable non-TLS ports by default. It is however recommended that non-TLS ports not be used unless running in a secure VPC or cluster. The certificates can be provided to the cluster using K8S secrets. Replace the template sections below with your Base64 encoded secret files.
If you skip this step, the Apica Ascent server automatically generates a ca and a pair of client and server certificates for you to use. you can get them from the ingest server pods under the folder /flash/certs
Save the secret file e.g. logiq-certs.yaml
. Proceed to install the secret in the same namespace where you want to deploy Apica Ascent
The secret can now be passed into the Apica Ascent deployment
HELM Option | Description | Defaults |
---|---|---|
| TLS certificate key pair and CA cert for TLS transport | No default |
3.4 Changing the storage class
If you are planning on using a specific storage class for your volumes, you can customize it for the Apica Ascent deployment. By default, Apica Ascent uses the standard
storage class
It is quite possible that your environment may use a different storage class name for the provisioner. In that case please use the appropriate storage class name. E.g. if a user creates a storage class ebs-volume
for the EBS provisioner for their cluster, you can use ebs-volume
instead of gp2
as suggested below
Cloud Provider | K8S StorageClassName | Default Provisioner |
---|---|---|
AWS | gp3 | EBS |
Azure | UltraSSD_LRS | Azure Ultra disk |
GCP | standard | pd-standard |
Digital Ocean | do-block-storage | Block Storage Volume |
Oracle | oci | Block Volume |
Microk8s | microk8s-hostpath |
Provisioning GP3 CSI Driver on AWS EKS - https://docs.logiq.ai/deploying-logiq/logiq-paas-deployment/deploying-logiq-eks-on-aws-using-cloudformation#5.3-enable-gp3-storage-class-for-eks
3.5 Using external AWS RDS Postgres database instance
To use external AWS RDS Postgres database for your Apica Ascent deployment, execute the following command.
HELM Option | Description | Default |
---|---|---|
| Deploy Postgres which is needed for Apica Ascent metadata. Set this to false if an external Postgres cluster is being used | true |
| Host IP/DNS for external Postgres | postgres |
| Postgres admin user | postgres |
| Postgres admin user password | postgres |
| Host Port for external Postgres | 5432 |
While configuring RDS, create a new parameter group that sets autoVaccum to true or the value "1", associate this parameter group to your RDS instance.
Auto vacuum automates the execution of VACUUM
and ANALYZE
(to gather statistics) commands. Auto vacuum checks for bloated tables in the database and reclaims the space for reuse.
3.6 Upload Apica Ascent professional license
The deployment described above offers 30 days trial license. Send an e-mail to support@apica.io
to obtain a professional license. After obtaining the license, use the apicactl tool to apply the license to the deployment. Please refer to apicactl
details at https://github.com/ApicaSystem/apicactl. You will need API-token from Apica Ascent UI as shown below
3.7 Customize Admin account
HELM Option | Description | Default |
---|---|---|
| Apica Ascent Administrator name | flash-admin@foo.com |
| Apica Ascent Administrator password | flash-password |
| Apica Ascent Administrator e-mail | flash-admin@foo.com |
3.8 Using external Redis instance
To use external Redis for your Apica Ascent deployment, execute the following command.
NOTE: At this time Apica Ascent only supports connecting to a Redis cluster in a local VPC without authentication. If you are using an AWS Elasticache instance, do not turn on encryption-in-transit or cluster mode.
HELM Option | Description | Default |
---|---|---|
| Deploy Redis which is needed for log tailing. Set this to false if an external Redis cluster is being used | true |
| Host IP/DNS of the external Redis cluster | redis-master |
| Host Port where external Redis service is exposed | 6379 |
3.9 Configuring cluster id
When deploying Apica Ascent, configure the cluster id to monitor your own Apica Ascent deployment. For details about the cluster_id
refer to section Managing multiple K8S clusters
HELM Option | Description | Default |
---|---|---|
global.environment.cluster_id | Cluster Id being used for the K8S cluster running Apica Ascent. See Section on Managing multiple K8S clusters for more details. | Apica AscentQ |
3.10 Sizing your Apica Ascent cluster
When deploying Apica Ascent, size your infrastructure to provide appropriate VCPU and memory requirements. We recommend the following minimum size for small. medium and large cluster specification from Section 1.3 values yaml files.
Apica Ascent Cluster | vCPU | Memory | NodeCount |
---|---|---|---|
small | 24 | 32 gb | 3 |
medium | 40 | 64 gb | 5 |
large | 64 | 128 gb | 8 |
3.11 NodePort/ClusterIP/LoadBalancer
The service type configurations are exposed in values.yaml as below
For e.g. if you are running on bare-metal and want an external LB to front Apica Ascent, configure all services as NodePort
3.12 Using Node Selectors
The Apica Ascent stack deployment can be optimized using node labels and node selectors to place various components of the stack optimally
The node label logiq.ai/node
above can be used to control the placement of ingest pods for log data into ingest optimized nodes. This allows for managing cost and instance sizing effectively.
The various nodeSelectors are defined in the globals section of the values.yaml file
In the example above, there are different node pools being used - ingest
, common
, db
, cache
and sync
Node selectors are enabled by setting enabled
to true
for globals.nodeSelectors
3.13 Installing Grafana
The Apica Ascent stack includes Grafana as part of the deployment as an optional component. To enable Grafana in your cluster, follow the steps below
The Grafana instance is exposed at port 3000 on the ingress controller. The deployed Grafana instance uses the same credentials as the Apica Ascent UI
3.14 Configuring ALB Ingress on EKS
Apica Ascent creates an Ingress resource in the namespace it is deployed.
Creating an OIDC provider for your EKS cluster - https://docs.aws.amazon.com/eks/latest/userguide/enable-iam-roles-for-service-accounts.htm
Please refer to the EKS configuration on how to automatically provision an ALB here - https://docs.aws.amazon.com/eks/latest/userguide/alb-ingress.html
4 Teardown
If and when you want to decommission the installation using the following commands
If you followed the installation steps in section 3.1 - Using an AWS S3 bucket, you may want to delete the s3 bucket that was specified at deployment time.
Last updated