| Feature | Shared | Old Dedicated plan | New Dedicated plan |
|---|---|---|---|
| Cluster size | 3 | 1, 3, 5, 7 | 1, 3, 4, 5, 6, 7, 8, 9 |
| Kafka Versions | 2.8.2 | 2.1.0 - 2.8.2 | 2.8.2 - 3.3.1 |
| Memory (RAM) | Fixed per plan | Adjustable (2gb-64gb) | |
| Storage | Fixed size | Custom size | |
| Email Support | |||
| Certificate based authentication | |||
| Centralized logs (Papertrail, Splunk, Loggly, Logentries etc) | |||
| Service Level Agreement (99.95%) | |||
| VPC Peering | |||
| Kafka browser | |||
| Manage Kafka Configuration (UI and API) | |||
| Manage Kafka Users and ACL (UI and API) | |||
| Manage Kafka Topics (UI and API) | |||
| Kafka specific Metrics | |||
| Kafka log view in browser | |||
| Multi-zone HA (3 nodes and higher) | |||
| Cluster downgrades | |||
| Seamless Upgrades (Kafka version & plan) | |||
| Increase disk size | |||
| Phone support |
Besides the comparison made in the table above there are some differences made to the technical side as well.
Always authenticate to the clusterOn the new plan both Zookeeper and Kafka are always protected with username and password, so even if you connect over a VPC peering you need to authenticate. This will allow you to manage access rights better but also if there were to sneak in an attacker in any of the peered VPCs they cannot access Kafka nor Zookeeper.
DNS managementOn the new dedicated plan we have changed how the DNS failover happens, previously we updated the cluster hostname to point to a single broker and updated this when it went offline but because of how DNS works this meant that your clients couldn't find an online broker for at least 30 seconds. Now we let the client know about all brokers so it can make a faster decision on where to continue sending its request.
New faster instance typesAll cloud providers keep releasing new instance types that are faster, the new plans is running on the latest possible instance on all clouds and has been built to allow us to faster upgrade when new ones are released.
Up to date operating system version and Java runtime versionThe new setup allows us to make sure that all servers run the latest and more secure version of both the operating system and the Java runtime.
All servers are continuously monitored for availability and performance. Most often we detect and address issues before they become a problem. If there is a problem with an instance it's automatically removed from the DNS load-balancer.
All plans are billed by the second, so you can try out even the largest instance types for mere pennies. Billing occurs at the end of each month, and you're only charged for the time an instance has been available to you.
We answer emails (support@cloudkarafka.com) 24/7 and for larger plans we also provide phone support. Please email us if you need help to get started or if you want to setup a call to discuss how CloudKarafka can support your business.
Make use of a large set of libraries to communicate with your CloudKarafka instance. Getting started example can be found in the documentation.
Our service-level agreement (SLA) guarantees at least 99.95% availability on all Commercial and Enterprise plans. CloudKarafka will refund 50% of the cost for outages longer than that. Request of refund must be submitted in writing, within 30 days from the outage to which they refer, via email to contact@cloudkarafka.com