Tyk OSS in Kubernetes with more ReplicaCount

Product FAQ
1

I am running Tyk OSS currently into my Kubernetes Cluster using the Tyk-headless helm chart , it seems working with me so far but while reading the documentation i noticed that i can not create multiple instances of Tyk POD within its Deployment as stated in this document Tyk Helm Chart

So the question now is : it is kind of limitation running OSS version ? or how can we create more instances of Tyk OSS ?

Thanks,
Aly

2

Hi,

The difficulty is that the deployment load balancer will stop you being able to communicate directly with each gateway to configure it. You could get around this by opening a shell in each gateway pod and configuring the api and policy definitions via the localhost interface then calling the reload API.

The other option would be to have multiple namespaces with individual gateways in each. That would allow you to use the operator in each namespace to configure each gateway, and then have another higher level load balancer to spread the traffic between gateways.

But don’t underestimate the amount of traffic that a single gateway can handle. It’s CPU bound so if provided with enough CPU an single gateway can handle a lot of traffic. Please see Planning for Production

Cheers,
Pete

3

Hi @AlyRagab, and welcome to the community.

Just to add to what @Pete has said, is there a reason you need more than one gateway?

Additionally, you could simply upgrade to Tyk pro which has the dashboard to handle this. The dashboard also handles syncing of the APIs across the gateway