New RunLLM response generated: >>>*# Question* Hi...
# runllm-flyte-trial
j
New RunLLM response generated: >>>*# Question* Hi Everyone. Could you tell me if it is safe to scale out
flyte-binary
deployment by increasing the number of pods in the deployment? Or if we want scaling out, we need to migrate to using
flyte-core
and separate deployments for propeller, datacatalog and etc? # Answer Scaling out the
flyte-binary
deployment by increasing the number of pods is not recommended. According to Flyte community discussions, simply adding more binary pods will not effectively scale the system. Instead, you should use
flyte-core
for better scalability, which allows for separate deployments of FlyteAdmin, FlytePropeller, DataCatalog, etc. For large-scale deployments,
flyte-core
is designed to handle more than a thousand concurrent workflows by deploying multiple replicas of each component and enabling independent scaling of FlytePropeller instances (Flyte GitHub issue). Would you like more details on how to migrate to
flyte-core
and configure it for scaling? Sources: - Flyte community discussions - Flyte GitHub issue Slack Link https://.slack.com/archives/CP2HDHKE1/p1727075325.456639 👍 0 button 👎 0 button 💡 Hint: Mention @runllm in the thread for followups.