Ma77Ball opened a new issue, #4291:
URL: https://github.com/apache/texera/issues/4291

   <h2>Problem</h2>
   <p>Currently, the default <code>values.yaml</code> does not specify resource 
requests or limits for most Texera microservices (webserver, 
workflow-computing-unit-manager, workflow-compiling-service, file-service, 
config-service, access-control-service), and the PostgreSQL and MinIO resource 
requests are set relatively high. This makes it difficult to run Texera on 
resource-constrained environments such as local development clusters (e.g., k3s 
on a laptop, Minikube, or small cloud VMs).</p>
   <p>The <code>values-development.yaml</code> overlay already addresses this 
by adding conservative resource constraints, but the base 
<code>values.yaml</code> should document sensible defaults so that new users 
aren't surprised by unbounded resource consumption.</p>
   <h2>Proposed Changes</h2>
   <ol>
   <li><strong>Reduce PostgreSQL resource requests/limits</strong> — lower CPU 
request from <code>1</code> to <code>0.25</code> and memory from 
<code>1Gi</code> to <code>256Mi</code> (with explicit limits).</li>
   <li><strong>Add MinIO resource limits</strong> — set memory request and 
limit to <code>256Mi</code>.</li>
   <li><strong>Add resource requests and limits for all Texera 
microservices:</strong></li>
   </ol>
   
   Service | CPU Request | CPU Limit | Memory Request | Memory Limit
   -- | -- | -- | -- | --
   webserver | 10m | 1000m | 256Mi | 1Gi
   workflow-computing-unit-manager | 10m | 1000m | 256Mi | 256Mi
   workflow-compiling-service | 10m | 1000m | 256Mi | 256Mi
   file-service | 10m | 1000m | 256Mi | 512Mi
   config-service | 10m | 1000m | 256Mi | 256Mi
   access-control-service | 10m | 1000m | 256Mi | 256Mi
   
   
   <h2>Motivation</h2>
   <ul>
   <li>Prevents pods from consuming unbounded resources and starving other 
services on a shared node.</li>
   <li>Enables Texera to run on machines with as little as ~8GB RAM.</li>
   <li>Helps Kubernetes schedule pods more effectively by providing the 
scheduler with accurate resource expectations.</li>
   <li>Provides a documented starting point that users can tune for their own 
hardware.</li>
   </ul></body></html>


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to