Documentation improvement

[zamazan4ik]

Hi!

After reading the documentation on https://www.fluvio.io/docs/ I have a bunch of suggestions for the documentation.

• Please add a search field to your documentation. It simplifies navigation in the documentation.
• Add documentation about supported hardware architectures. E.g. about supported architectures for different operating systems, some specific requirements to the supported instructions, if you have any (e.g. maybe AVX is required on x86-64 - I do not know). Perhaps some kind of table "OS-Arch" will be better. Right now it's unclear what architectures for that operating system are supported.

• Does it support zero-downtime upgrades? If yes, how it could be performed? If not, could you please mention it explicitly in the documentation? And, possibly, provide a scheme for how zero-downtime could be achieved (like a two-cluster scheme with Cross-Cluster Replication or smth like that)
• Documentation says Fluvio does not support downgrade. So, if after the upgrade I will got an error - what should I do? Would be great to see an emergency plan for that (like restoring the previous cluster from a backup)
• Does it support any kind of multi-cluster support out of the box? E.g. Cross-Cluster Replication, Stand-By Clusters, etc.
• I suggest removing these benchmarks from the documentation since they are not related at all to the project itself and could be confusing. Would be much better if you could perform some comparison to Kafka/RabbitMQ/NATS/Pulsar/etc. and make these benchmarks public (publish source code) and reproducible. This would help a lot during the in-house evaluation of your product.
• Also, if you care about performance, you could think about a thing like that: https://benchmark.clickhouse.com/. One more idea - try to apply PGO to Fluvio. If it helps (it definitely could - link) - document it somewhere in the documentation.
• How could I secure my cluster? Users support, roles, permissions, authentication methods, and how it could be configured.
• Is there any quotas support? E.g. specify request quotas per user and deny requests if the quota is exceeded.
• Is there any Prometheus/Zabbix/etc integration for monitoring purposes?
• Are there any ready-to-use Grafana dashboards for monitoring?
• Provide reference architectures for different deployments (single-node, multiple-node). Maybe, based on some public cloud providers like AWS/Azure/GCP.
• It's not completely clear, does Fluvio use any kind of "cluster" or not. But if yes, I want to see some cluster design recommendations: minimal node count (for avoiding split-brain problem if it's applicable to the application), latency requirements between nodes, notes about geo-replicated clusters (is it possible to configure so, how, what about network latencies and not-so-stable network connections, etc)
• How to backup and restore Fluvio cluster?
• Would be great to see Jepsen-like tests for Fluvio
• Would be great to have your roadmap public. It helps with estimating when important for the users features will be available.
• Do you have the plan to support something like Kafka's "exactly-once" semantic?
• Does it have Dead-Letter Queue (DLQ) support?
• (Just to highlight) Please revive the issue about zstd compression support :)
• (Just a nit :) MacOS -> macOS

This list could be used as a documentation epic - I hope it will be helpful for that purpose.

Thanks in advance.

[drc-infinyon]

@zamazan4ik - Thanks for this excellent issue. This is so well detailed.
@tjtelan - Let's connect next week to discuss this documentation feedback and figure out how we can clean up.

[sehz]

Thanks for great detailed feedback!

[github-actions]

Stale issue message