THE FIGURE SHORTCODE
Hugo has figure shortcode built-in, so you can easily add figure captions or hyperlink rel attributes to images. Documentations can be found here:
CONFLUENT CLOUD KAFKA VS GOOGLE CLOUD PUBSUB FEATURE COMPARE 2020
| Feature | Confluent Cloud Kafka | Google Cloud Pubsub | Notes |
|---|---|---|---|
| Data Retention | Set retention per topic in Confluent Cloud, including unlimited retention with log compaction. | Retains unacknowledged messages in persistent storage for 7 days from the moment of publication. There is no limit on the number of retained messages. Have to write custom subscriber/publisher to save beyond 7 days [L] + ongoing BAU [S] | |
| Replay | A consumer request an “offset”, however the retention period is dictated by the broker config | “Snapshots” can be created for later replay by these are limited to 7 days as per retention policy. As per above, a custom subscriber/publisher to save/replay message. [L] + ongoing BAU [S] | |
| Message Ordering | Yes, within a partition, In general, messages are written to the broker in the same order that they are received by the producer client. | No, Pub/Sub provides a highly-available, scalable message delivery service. The tradeoff for having these properties is that the order in which messages are received by subscribers is not guaranteed. While the lack of ordering may sound burdensome, there are very few use cases that actually require strict ordering. | |
| Delivery Semantics | Exactly-once delivery semantics | At-least-once, exactly once possible with dataflow | |
| Latency | Advertised at being able to “Achieve sub 30 ms latency at scale”, no mention of this in SLA. | Pub/Sub does not guarantee message delivery latency | |
| Uptime | “Service Level” Monthly Uptime Percentage of at least 99.95%.Is this 99.95% of GCP’s 99.95%? | “Service Level Objective” Monthly Uptime Percentage to Customer of at least 99.95% | Both offer credits to account if not met but this is unlikely to be suitable for an enterprise org |
| Schema Registry | Yes for Avro schemas and very new (GA August 2019) | No, Data Catalog is in beta and could be used to build one [XL] | |
| IAM / ACL | “Preview” for Role/SAML, ACLs You are provided auth keys that you need to store/share/rotate. These could be stored in Cloud KMS, however, this would need to be automated [XL] + ongoing BAU [M] | Standard Google IAM | |
| Encryption | Yes in transit and at rest with NO payload encryption. Clients are responsible for writing custom encryption/decryption connecting to (e.g) Cloud KMS. A custom library would need to be written that is used by everyone for publishing and subcribing [L] | Cloud KMS(HSM/Software/BYOK/External Key Manager) with CMEK | |
| VPC Security | Unknown/No, can Confluent Cloud be made to respect VPC service controls? | VPC Service Controls protection applies to all push and pull operations except existing Pub/Sub push subscriptions | |
| Stream Processing | “fully-managed KSQL”, no kafka streams, would have to run Kafka Streams/Storm cluster connecting to Confleunt Cloud which is likely to introduce latency | Apache Beam / Cloud Dataflow fully managed | |
| Costs per 130GB | $37 | $39 | Based on example calc on confluent and google pricing calc |
| Priority Queues | Yes | No, but can segregate by topic | |
| Multi-zone high availability | Not advertised “Contact Confluent” | Yes |
MULTI LANGUAGE BEAM PIPELINES
Apache Beam is an open-source unified programming model and framework for defining and executing big data processing pipelines. It provides a way to write data processing code that is portable across different execution engines or runtimes, such as Apache Flink, Apache Spark, Google Cloud Dataflow, and more.
TERRAFORM CLOUD DEVELOPMENT KIT
Terraform’s Cloud Development Kit (CDK) let’s you use other languages to define your cloud infra.
MISCONFIGURED KUBEFLOW WORKLOADS ARE A SECURITY RISK
“During April, we observed deployment of a suspect image from a public repository on many different clusters. The image is ddsfdfsaadfs/dfsdf:99. By inspecting the image’s layers, we can see that this image runs an XMRIG miner:” Source
DATAFLOW REAL TIME + AGGREGATE
A great way to split up your pipeline based on the urgency of results aggregate-data-with-dataflow